Search

[Display omitted] The nanoporous polydimethylsiloxane (PDMS) surfaces of a rectangular microfluidic channel, selectively uptakes water molecules, concentrating the solute molecules in an aqueous phase, that could drive phase transitions. Factors such as surface wettability, channel geometry, the surface-to-volume ratio, and surface topography of the confinements could play a key role in tuning the phase transitions spatio-temporally. Using a lyotropic chromonic liquid crystal as model biological material, confined within nanoporous microfluidic environments, we study molecular assembly driven by nanoporous substrates. By combining timelapse polarized imaging, quantitative image processing, and a simple mathematical model, we analyze the phase transitions and construct a master diagram capturing the role of surface wettability, channel geometry and embedded topography on programmable lyotropic phase transitions. Intrinsic PDMS nanoporosity and confinement cross-section, together with the imposed wettability regulate the rate of the N -M phase transition; whereas the microfluidic geometry and embedded topography enable phase transition at targeted locations. We harness the emergent long-range order during N -M transition to actuate elasto-advective transport of embedded micro-cargo, demonstrating particle manipulation concepts governed by tunable phase transitions. Our results present a programmable physical route to material assembly in microfluidic environment, and offer a new paradigm for assembling genetic components, biological cargo, and minimal synthetic cells. [ABSTRACT FROM AUTHOR]

Sobel is one of the most popular edge detection operators used in image processing. To date, most users utilize the two-directional 3×3 Sobel operator as detectors because of its low computational cost and reasonable performance. Simultaneously, many studies have been conducted on using large multi-directional Sobel operators to satisfy their needs considering the high stability, but at an expense of speed. This paper proposes a fast graphics processing unit (GPU) kernel for the four-directional 5×5 Sobel operator. To improve kernel performance, we implement the kernel based on warp-level primitives, which can significantly reduce the number of memory accesses. In addition, we introduce the prefetching mechanism and operator transformation into the kernel to significantly reduce the computational complexity and data transmission latency. Compared with the OpenCV-GPU library, our kernel shows high performances of 6.7x speedup on a Jetson AGX Xavier GPU and 13x on a GTX 1650Ti GPU. • Implemented an efficient GPU kernel of a four-directional 5x5 Sobel operator. • Implemented GPU kernel based on warp-level primitives. • Hid latency caused by data transmission using the prefetching mechanism. • Accelerated the computation of diagonal directions using a two-step optimization approach. • Achieved an extremely high processing speed, which is almost 13x faster than that of the OpenCV-GPU library. [ABSTRACT FROM AUTHOR]

We apply methods of machine-learning, such as neural networks, manifold learning and image processing, in order to study 2-dimensional amoebae in algebraic geometry and string theory. With the help of embedding manifold projection, we recover complicated conditions obtained from so-called lopsidedness. For certain cases it could even reach ∼ 99 % accuracy, in particular for the lopsided amoeba of F 0 with positive coefficients which we place primary focus. Using weights and biases, we also find good approximations to determine the genus for an amoeba at lower computational cost. In general, the models could easily predict the genus with over 90% accuracies. With similar techniques, we also investigate the membership problem, and image processing of the amoebae directly. [ABSTRACT FROM AUTHOR]

[Display omitted] • Development of an experimental platform based on a smart resealable microfluidic chip coupled to an inverted optical fluorescence microscope and a high-speed camera to perform in situ monitoring of the assembly of photoluminescent nano-objects on surfaces. • Deciphering the 3D assembly mechanisms of CdSe/CdZnS colloidal quantum nanoplatelets by nanoxerography, coupling experiments with numerical simulations. • Tuning on demand the thickness of CdSe/CdZnS quantum nanoplatelet-based patterns from monolayers to multilayers. • Realization of an array of 3 µmx3 µmx60 nm CdSe/CdZnS quantum nanoplatelets based pixels on a cm2 transparent surface area towards applications in photoemissive displays. Hypothesis: Due to their unique quantum yield and photostability performances, quantum nanoplatelets are very promising building blocks for future generations of displays. The directed assembly of such colloidal nano-objects in the shape of micro-pixels is thus the next mandatory step to reach this goal. Selectively trapping them on electrostatically charged patterns by nanoxerography could be a versatile and appealing strategy but requires a full understanding of the assembly mechanisms in order to make the most of their integration. Experiments: We propose an experimental platform based on a smart resealable microfluidic chip coupled to an inverted optical fluorescence microscope and a high-speed camera for in situ access of such assembly mechanisms, using CdSe/CdZnS quantum nanoplatelets as model nano-objects. The photoluminescence signal of the nanoplatelet patterns is thus recorded in real time during their assembly and data extracted after image processing. Findings: The coupling of experimental results and numerical simulations evidences the main role of advection at the origin of this directed nanoparticle trapping. Deep understanding of the involved mechanisms and tuning of experimental parameters allow to make high resolution quantum nanoplatelet based micro-pixels with a fine control of their lateral and vertical dimensions. [ABSTRACT FROM AUTHOR]

[Display omitted] Contact angle goniometry suffers from inherent optical challenges such as scattering and diffraction near the triple contact line (TCL) rendering erroneous results. Alternatively, the cantilever-based direct adhesion measurement was constrained to low-energy surfaces to date due to the inability of the probe droplet to retract (pull-off) from high-energy surfaces completely. The present study revisits the cantilever approach from a fundamental physical perspective and generalizes the approach to render it wettability invariant. The adhesive wetting interaction between a probe droplet (attached to a cantilever) with the test substrate is recorded with a high-speed camera. Image processing and subsequent motion analysis enable us to accurately calculate the adhesion force (in the sub-micron range) exhibited by the test substrate. We experimentally demonstrate the contact line depinning (and not the droplet pull-off) to be the only prerequisite for accurately quantifying the characteristic adhesion force. We also reveal that depinning precedes the onset of cantilever retraction due to the inertia effect. Further, we establish that the characteristic adhesion corresponds to zero acceleration of the cantilever and not to its maximum deflection. The inferences of the study will be beneficial in the rational design of direct wetting characterization methods for any substrate. [ABSTRACT FROM AUTHOR]

• Review comprises 67 methods for quantifying PVS applied in 60 studies up to September 2023. • Morphological and Hessian filters, and U-Net configurations are the most widely applied approaches. • PVS burden is associated with age, sex, hypertension, diabetes, WMH, sleep and cognition. • Occupation, ethnicity and genetic traits also relate to PVS burden. Research into magnetic resonance imaging (MRI)-visible perivascular spaces (PVS) has recently increased, as results from studies in different diseases and populations are cementing their association with sleep, disease phenotypes, and overall health indicators. With the establishment of worldwide consortia and the availability of large databases, computational methods that allow to automatically process all this wealth of information are becoming increasingly relevant. Several computational approaches have been proposed to assess PVS from MRI, and efforts have been made to summarise and appraise the most widely applied ones. We systematically reviewed and meta-analysed all publications available up to September 2023 describing the development, improvement, or application of computational PVS quantification methods from MRI. We analysed 67 approaches and 60 applications of their implementation, from 112 publications. The two most widely applied were the use of a morphological filter to enhance PVS-like structures, with Frangi being the choice preferred by most, and the use of a U-Net configuration with or without residual connections. Older adults or population studies comprising adults from 18 years old onwards were, overall, more frequent than studies using clinical samples. PVS were mainly assessed from T2-weighted MRI acquired in 1.5T and/or 3T scanners, although combinations using it with T1-weighted and FLAIR images were also abundant. Common associations researched included age, sex, hypertension, diabetes, white matter hyperintensities, sleep and cognition, with occupation-related, ethnicity, and genetic/hereditable traits being also explored. Despite promising improvements to overcome barriers such as noise and differentiation from other confounds, a need for joined efforts for a wider testing and increasing availability of the most promising methods is now paramount. [ABSTRACT FROM AUTHOR]

The feasibility of using VIS-NIR and SWIR hyperspectral imaging (HSI) to determine the salinity, soluble solids, and water content in salted kimchi cabbages during the salting process was assessed. Hyperspectral images of kimchi cabbages acquired at different salt concentrations and periods of the salting process were developed as prediction and discriminant models for the qualitative properties. Principal component analysis showed that the distribution according to the spectral characteristics of kimchi cabbages could be grouped, and data reduction for optimal model developing was attempted based on the key wavelengths selected in the loading plot. The optimal partial least squares regression models for the prediction of salinity and soluble solids had best performance by achieving R2 p values of over 0.86 and 0.90 in the VIS-NIR and SWIR regions, respectively. The accuracy and specificity of the optimal PLS-DA model for salinity level have improved efficiency to over 0.93. Visualization of the spectrum information in each pixel of the hyperspectral image using the PCA model displayed the salinity level in kimchi cabbages under different salting conditions. The prediction models were sufficiently accurate to consider HSI as a useful tool for controlling and optimizing the salting process. [Display omitted] • First time hyperspectral system was used to monitor KC salting process. • Predictive abilities of PLS models were compared in both VIS-NIR and SWIR. • The salting level discrimination models for KC were established with good accuracy. • Optimal wavelengths were selected to develop multispectral imaging system. • Salting level distribution maps was visualized by PCA image processing algorithm. [ABSTRACT FROM AUTHOR]

The analysis of high resolution whole slide tissue images is a computationally expensive task, which adversely impacts effective use of pathology imaging data in research. We propose runtime solutions to enable efficient execution of pathology image analysis applications on modern distributed memory hybrid platforms equipped with both CPUs and GPUs. Hybrid systems offer significant computation capacity, but taking advantage of this computing power is complex. An application developer may have to implement multiple versions of data processing codes targeted for different computing devices. The developer also has to tackle the challenges of efficiently distributing computational load among the nodes of a distributed memory machine and among computing devices on a node. This is particularly difficult in analysis of high resolution images because of irregular computing costs of processing different image regions. In order to address these problems, we have leveraged a high-level image processing language (Halide) and integrated it into our runtime system called Region Templates (RT). The language simplifies the application development while generating code for multiple devices, such as CPU and GPU. The integration with RT allows for efficient multiple node hybrid execution. We also developed a novel cost-aware data partitioning (CADP) strategy that considers the workload irregularity to minimize load imbalance. Our experimental evaluation shows significant performance improvements on hybrid CPU-GPU machines, as compared with using a single processor (CPU or GPU), as well as on multi-GPU systems. CADP resulted in 1.7× better performance than other workload partitioning approaches (e.g., KD-Trees) on a hybrid machine and was up to 2.24× faster in multi-node settings. • Pathology image analysis is computed intensive due to the high-resolution images. • We evaluate the use of Halide with Region Templates (RT) to speedup these analyses. • We support execution on machines with CPU-GPU and unified programing interface. • We propose a novel cost-aware data partitioning strategy targeting hybrid machines. • Our solutions attained better performance than competitor partitioning strategies. [ABSTRACT FROM AUTHOR]

• IDCT and FDCT written in Halide. • 200 Lines of Halide replace over 20,000 lines of Assembly Language. • 5-25% faster for decoding. • 10-40% faster for encoding. • Significantly easier to maintain and port to new architectures. The Discrete Cosine Transform (DCT) is commonly used for image and video coding and very efficient implementations of the forward and inverse transforms are of great importance. The popular libjpeg-turbo library contains handwritten, highly-optimised assembly language DCT implementations utilizing SIMD instruction sets for a variety of architectures. We present an alternative approach, implementing the 8x8 IDCT and FDCT written in the functional image processing language Halide. We show how less than 200 lines of Halide can replace over 20,000 lines of code the libjpeg-turbo library to perform JPEG encoding and decoding. The Halide implementation is compared for ARMv8 NEON and x86-64 SIMD extensions and shows a 5–25 percent performance improvement over the SIMD code in libjpeg-turbo for decoding and a 10–40 percent improvement for encoding. The Halide code is significantly easier to maintain and port to new architectures than the existing code. [ABSTRACT FROM AUTHOR]

Microplastics (MPs), the emerging pollutants appeared in water environment, have grabbed significant attention from researchers. The quantitative method of spherical MPs is the premise and key for the study of MPs in laboratory researches. However, the manual counting is time-consuming, and the existing semi-automated analysis lacked of robustness. In this study, a highly accurate quantification method for spherical MPs, called VS120-MC was proposed. VS120-MC consisted of the digital slide scanner VS120 and the MPs image processing software, MPs-Counter. The full-area scanning photography was employed to fundamentally avoid the error caused by random or partition sampling modes. To accomplish high-performance batch recognition, the Weak-Circle Elimination Algorithm (WEA) and the Variable Coefficient Threshold (VCT) was developed. Finally, lower than 0.6% recognition error rate of simulated images with different aggregated indices was achieved by MPs-Counter with fast processing speed (about 2 s/image). The smallest size for VS120-MC to detect was 1 μm. And the applicability of VS120-MC in real water body was investigated. The measured value of 1 μm spherical MPs in ultra-pure water and two kinds of polluted water after digestion showed a good linear relationship with the Manual measurements (R2 = 0.982,0.987 and 0.978, respectively). For 10 μm spherical MPs, R2 reached 0.988 for ultra-pure water and 0.984 for both of the polluted water. MPs-Counter also showed robustness when using the same set of parameters processing the images with different conditions. Overall, VS120-MC eliminated the error caused by traditional photography and realized an accurate, efficient, stable image processing tool, providing a reliable alternative for the quantification of spherical MPs. • Full-area photography and semi-automatic image recognition of spherical microplastics were implemented. • Accurate quantification of spherical microplastics in both clean or polluted water for the size limit of 1 μm. • Efficient and stable batch process of microplastics images can be achieved by the homemade software MPs-Counter. • The average processing time for MPs-Counter was 2 s/image, saving about 10 times of manual counting. [ABSTRACT FROM AUTHOR]

Over the past few years, GPUs have found widespread adoption in many scientific domains, offering notable performance and energy efficiency advantages compared to CPUs. However, optimizing GPU high-performance kernels poses challenges given the complexities of GPU architectures and programming models. Moreover, current GPU development tools provide few high-level suggestions and overlook the underlying hardware. Here we present Starlight, an open-source, highly flexible tool for enhancing GPU kernel analysis and optimization. Starlight autonomously describes Roofline Models, examines performance metrics, and correlates these insights with GPU architectural bottlenecks. Additionally, Starlight predicts potential performance enhancements before altering the source code. We demonstrate its efficacy by applying it to literature genomics and physics applications, attaining speedups from 1.1× to 2.5× over state-of-the-art baselines. Furthermore, Starlight supports the development of new GPU kernels, which we exemplify through an image processing application, showing speedups of 12.7× and 140× when compared against state-of-the-art FPGA- and GPU-based solutions. • We enrich the incomplete information provided by NVIDIA profilers. • Starlight can support the development of an application from the ground up. • Starlight predicts potential performance enhancements before altering the source code. • Automatic Roofline Model generation for any CUDA-capable GPU. • A qualitative overview of the various state-of-the-art solutions for GPU kernel optimization and Roofline Model generation. [ABSTRACT FROM AUTHOR]

We study a nonlinear reaction-convection equation with a degenerate diffusion of Perona-Malik's type and a monostable reaction term. Under quite general assumptions, we show the presence of wavefront solutions and prove their main properties. In particular, such wavefronts exist for every speed in a closed half-line and we give estimates of the threshold speed. The wavefront profiles are also strictly monotone and their slopes are uniformly bounded by the critical values of the diffusion. [ABSTRACT FROM AUTHOR]

Copyright infringement adversely affects the interest of copyright holders of images and videos which are uploaded to different websites and peer-to-peer image sharing systems. This paper addresses the problem of detecting copyright infringement so that copyright holders are given due credit for their work. There are several images and videos that are shared every day by millions of users with some amount of modification in images and videos originally uploaded by the copyright holders such as photographers, graphic designers, and video providers. Copyright violators, who are not the original creators of multimedia content modify them using image processing and frame modification techniques such as grayscale conversion, cropping, rotation, frame compression, and frame speed manipulation. Then, upload the tampered images and videos. To address this problem, we propose an IPFS-based (InterPlanetary File System-based) decentralized peer-to-peer image and video sharing platform built on blockchain technology. We use a perceptual hash (pHash) technique to detect copyright violations of multimedia. When multimedia is to be uploaded to the IPFS, the pHash of the same content is determined and checked against existing pHash values in the blockchain network. Similarity with existing pHash values would result in the multimedia being detected as tampered with. Blockchain technology offers the advantage of non-involvement of a third party and consequently the avoidance of a single point of failure. • We present a platform of blockchain and IPFS to store multimedia objects as transactions. • The proposed framework ensures multimedia availability, immutability, and transparency. • The proposed approach is fully distributed where copyright information can be verified. [ABSTRACT FROM AUTHOR]

eX-Poro-HydroDynamic (XPHD) lubrication presents a different scientific approach to dealing with tribological problems. It is an innovative inter- and multidisciplinary research topic which offers a promising sliding solution for various applications, such as thrust bearings, various guide components and in terms of load capacity and damping. XPHD lubrication is a lubrication mechanism of biomimetic inspiration which features an additional parameter to the system "the porous media". It consists of self-sustained fluid films generated within compressible porous layers imbibed with liquids in replacement for using the fluid film only as in the classic lubrication system. Soft and porous structures imbibed with liquids generate a high load support under compression. The load support is generated through the resistance to flow inside the porous material. During compression, the resistance to flow and load support increases the greater the compression rate. The main objective of this work is then to understand the behavior of the fluid flow inside the porous structures when subjected to axial compression stress. In the scientific literature, the works studying the flow in compressible materials are essentially experimental because of their very complex geometrical shape, the CFD (Computational Fluid Dynamics) simulations offer an economical solution to study the performance of this new concept of lubrication. To create the geometry, the morphological structures of foam samples are reconstructed at different levels of compression rates from 3D (Three Dimensional) X-ray microtomography. This is achieved by using the commercial software Avizo that allows to process 3D images and create 3D meshes suitable for numerical simulations. The numerical simulations of flows will be performed with the solver IcoFoam of the toolbox OpenFOAM for incompressible laminar flows, making it possible to study the pressure drop in these porous structures. The performed simulations were made with a polyurethane foam of 96% porosity using five compression rates for creating the different structures. The analysis of the numerical simulations shows the impact of the polyurethane foam compression on different key parameters such as the decrease in the permeability as function of the compression rate, the anisotropy of the flow within the compressible structure and the actual increase in the tortuosity generated by the compression of the foam and the variation of the porosity. [ABSTRACT FROM AUTHOR]

Shadow detection is significant for scene understanding. Ambiguities in a shadow image, such as shadow-like non-shadow regions and shadow regions with non-shadow patterns, are still very challenging for prevalent CNN-based methods. This work attempts to alleviate this problem from a new perspective of shape semantics, and then proposes a Semantic-aware Transformer (SaT) in a multi-task learning manner. Concretely, we first propose a shadow detection network based on the recent progress of Transformer architecture, allowing us to capture significant global interactions between contexts. Next, we design a multi-task learning framework, combining shadow supervision and semantic supervision to perform a semantic-aware shadow detection. Finally, we introduce a simple yet effective information buffer unit to overcome the gradient signal conflict from multi-task learning. Experimental results on three public benchmark datasets (i.e., ISTD, SBU, and UCF) show that our SaT can effectively detect ambiguous cases and achieve state-of-the-art results. • We propose a shadow detection method using shape semantics. • We design an efficient shadow detection network based on Transformer. • We construct a multi-task learning framework combining shadow and semantic supervision. • An information buffer unit is proposed to coordinate different supervised tasks. [ABSTRACT FROM AUTHOR]

The impact of a coating solution containing chitosan, lignocellulosic nanofibers (LCNF), and raspberry leaf extract on strawberry quality was assessed. Strawberries were coated, inoculated with Aspergillus flavus (a major aflatoxin producer), and stored at 5 ± 1 °C for 16 days. An automated fruit quality evaluation system was developed, and daily images were taken for analysis. ImageJ software was used to evaluate colour, size loss, and fungal damage. Coating solution significantly reduced fruit size loss, with coated strawberries showing 18.4 ± 12.8% size loss after 16 days, compared to 30.2 ± 8.4% in the control group and 25 ± 10.7% in uncoated strawberries. Regarding colour, lower colour changes were observed in coated fruit compared to the control during storage (p ≤ 0.05). Fungal damage did not significantly differ between treatments during storage (p > 0.05). Furthermore, the study demonstrated that image analysis and visual estimation methods for assessing fruit fungal damage yielded similar results, with a strong correlation (R2 = 0.87). This suggests that the proposed image analysis methodology is effective for the assessment of fungal damage in strawberries compared to visual estimation. Chitosan-LCNF-raspberry extract coating solution shows promise in enhancing strawberry quality by reducing size loss during storage and managing fungal development. [Display omitted] • Coating effect was evaluated on strawberries inoculated with Aspergillus flavus. • Image analysis assessed fungal damage, size loss and colour. • Kinetic models were created to assess fruit quality and safety. • Chitosan-LCNF-raspberry extract coating extends strawberry shelf-life. • Coated strawberries showed lower size loss and similar fungal damage. [ABSTRACT FROM AUTHOR]

Attention mechanisms have become of crucial importance in deep learning in recent years. These non-local operations, which are similar to traditional patch-based methods in image processing, complement local convolutions. However, computing the full attention matrix is an expensive step with heavy memory and computational loads. These limitations curb network architectures and performances, in particular for the case of high resolution images. We propose an efficient attention layer based on the stochastic algorithm PatchMatch, which is used for determining approximate nearest neighbors. We refer to our proposed layer as a "Patch-based Stochastic Attention Layer" (PSAL). Furthermore, we propose different approaches, based on patch aggregation, to ensure the differentiability of PSAL, thus allowing end-to-end training of any network containing our layer. PSAL has a small memory footprint and can therefore scale to high resolution images. It maintains this footprint without sacrificing spatial precision and globality of the nearest neighbors, which means that it can be easily inserted in any level of a deep architecture, even in shallower levels. We demonstrate the usefulness of PSAL on several image editing tasks, such as image inpainting, guided image colorization, and single-image super-resolution. Our code is available at: https://github.com/ncherel/psal. • An efficient attention layer for images based on Approximate Nearest Neighbor Search. • Our layer has a log-linear computational complexity and linear memory complexity. • In contrast standard attention has quadratic computational and memory complexity. • We detail how differentiability is preserved despite the use of nearest neighbors. • Low memory costs permit to adapt SOTA editing algorithms to high-resolution images [ABSTRACT FROM AUTHOR]

Graph-tensor learning operations extend tensor operations by taking the graph structure into account, which have been applied to diverse domains such as image processing and machine learning. However, the running time of graph-tensor operations increases rapidly with the number of nodes and the dimension of data on nodes, making them impractical for real-time applications. In this paper, we propose a GPU library called cuGraph-Tensor for high-performance graph-tensor learning operations, which consists of eight key operations: graph shift (g-shift), graph Fourier transform (g-FT), inverse graph Fourier transform (inverse g-FT), graph filter (g-filter), graph convolution (g-convolution), graph-tensor product (g-product), graph-tensor SVD (g-SVD) and graph-tensor QR (g-QR). cuGraph-Tensor supports scalar, vector, and matrix data processing on each graph node. We propose optimization techniques on computing, memory accesses, and CPU–GPU communications that significantly improve the performance of the graph-tensor learning operations. Using the optimized operations, cuGraph-Tensor builds a graph data completion application for fast and accurate reconstruction of incomplete graph data. In the experiments, the proposed graph learning operations achieve up to 142. 12 × speedups versus CPU-based GSPBOX and CPU MATLAB implementations running on two Xeon CPUs. The graph data completion application achieves up to 174. 38 × speedups over the CPU MATLAB implementation, and up to 3. 82 × speedups with better accuracy over the GPU-based tensor completion in the cuTensor-tubal library. • We propose a GPU-based library of eight Graph-Tensor learning operations. • We propose optimization techniques on computing, memory access, and communications. • We develop a graph data completion application to demonstrate the library. • The proposed library achieves up to 142.12 × speedups over CPU-based implementations. • The application achieves up to 3.82 × speedups and better accuracy over GPU-based work. [ABSTRACT FROM AUTHOR]

Over the last decade, the sparse representation model has led to remarkable results in numerous signal and image processing applications. To incorporate the inherent structure of the data and account for the fact that not all support patterns are equally likely, this model was enriched by enforcing various structural sparsity patterns. One plausible such extension of classic sparse coding, instigated by the emergence of graph signal processing, is graph regularized sparse coding. This model explicitly considers the intrinsic geometrical structure of the data domain, and has been successfully employed in various applications. However, emphasis was given to developing algorithmic solutions, and to date, the theoretical foundations to this problem have been lagging behind. In this work, we fill this gap and present a novel theoretical analysis of the graph regularized sparse coding problem, providing worst-case guarantees for the stability of the obtained solution, as well as for the success of several pursuit techniques. Furthermore, we formulate the conditions for which the superiority of the graph regularized sparse coding solution over the structure-agnostic sparse coding counterpart is established. [ABSTRACT FROM AUTHOR]

• Spreading on the laser textured surface shows the formation of three spreading fronts. • The anisotropy of spreading for all types of liquid fronts was found. • Biologically relevant liquid imbibition into the nanotexture follows universal power law. The distribution of biological objects upon the spreading of biologically relevant dispersions over laser textured surfaces is affected by the dispersion composition and substrate chemistry and roughness. To examine the role of the substrate texture in biologically relevant liquid spreading, the dynamic behavior of droplets of water and dispersions of bacterial cells and viruses and dynamic behavior of droplet/air surface tension were addressed. A new procedure to simultaneously distinguish three different spreading fronts was developed. The study of spreading of water and the biologically relevant liquids over the laser textured substrate indicate the development of three spreading fronts associated with the movement of bulk droplet base, the flow along the microchannels, and the nanotexture impregnation. The anisotropy of spreading for all types of liquid fronts was found. Despite the expected difference in the rheological behavior of water and the biologically relevant liquids, the spreading of all tested liquids was successfully described by power-law fits. The droplet base spreading for all tested liquids followed the Tanner law. The advancing of water and dispersions in the microchannels along both fast and slow axes was described by Washburn type behavior. The impregnation of the nanotexture by water and biologically relevant liquids demonstrated universality in power fit description with an exponent n = 0.23. [ABSTRACT FROM AUTHOR]

We propose a method to measure the interfacial tension characterizing the interface between two immiscible liquids of practically the same density. In this method, a cylindrical liquid bridge made of one the liquids is vibrated laterally inside a tank filled with the other. The first resonance frequency is determined and equated to the first eigenfrequency of the m = 1 linear mode to infer the interfacial tension value. The method does not involve the density jump across the interface. Therefore, its accuracy is affected neither by the smallness of the Bond number nor by errors of the density difference. The experimental setup is relatively simple, and the procedure does not use image processing techniques. The results satisfactorily agree with those measured by TIFA-AI (Theoretical Fitting Image Analysis-Axisymmetric Interfaces) for the same liquid bridges when the density difference is sufficiently large for TIFA-AI to be valid. We conduct numerical simulations of the Navier-Stokes equations to determine the best parameter conditions for the proposed method. The transfer function characterizing the frequency response of the fluid configuration is measured in some experiments to quantify non-linear effects and to study the role played by the outer bath vibration. [ABSTRACT FROM AUTHOR]

Herein, a parallel implementation in OpenMP of the Image Block Representation (IBR) for binary imagesis investigated. The IBR is a region-based image representation scheme that represents the binary image as a set of non-overlapping rectangular areas with object level, called blocks. The IBR permits the execution of operations on image areas instead of image points and therefore leads to a substantial reduction of the required computational complexity. The experimental and the analytically derived results from parallel implementation in OpenMP, on a multicore computer, proved that a very good overall performance can be achieved. • The Image Block Representation (IBR) has been used for the creation of fast image processing and analysis algorithms. • A parallel implementation using OpenMP on multicore computers is investigated. • Theoretical analysis of the parallel algorithm's computational efficiency is given. • Significant speedups are achieved using 40 cores and the experimental results matched the theoretical predictions. [ABSTRACT FROM AUTHOR]

In this paper, two aims are sought. First, the path to a minimum-error colour calibration and its feasibility for non-invasive cultural heritage conservation is explored. Using RGB information as temporal cues of the conservation state of the pieces under monitoring, an adaptive calibration function, working in syntony with a colour calibration chart destined to general usage, has been proposed. Since the context involves citizen participation, the pictures to calibrate are crowdsourced and the cameras of origin unknown; therefore, the requirements of the work context assume images that imply a heavy colorimetric uncertainty whilst maintaining their structural content. Therefore,a differentiable, multidimensional, adaptive transfer curve that offers the best trade-off between calibration error (staying in minimum perceivable colour differences for tangible materials) and calculation requirements is refined. The conception of this system offers flexible possibilities of an effective cultural heritage conservation without heavy costs on money and human effort or relying on external solutions, while standing on par in quality with state-of-art technologies, achieving therefore a balance between generalist usage and specific solutions. Secondly, a neural network-based final process performs the final tuning of the calibration functions, and its conclusions shed light on how a calibration process works together with the colour chart on an algebraic level. The verification of this phenomenon opens possibilities of different paths on how to proceed with colour calibrations depending on the conditions of acquisition and the content of interest, so the minimum error can be always achieved regardless of the case. The ideas presented here are an expansion and conclusion of the developments and results previously published by the same authors. • A custom, reliable and adaptive colour calibration system of minimum error is feasible. • A tight calibration colour chart distribution constrains the function curve calculation. • Edge delimitation only defines a cultural heritage appropriate calibration chart. • A custom calibration framework is optimum for heritage preservation requirements. • Designing chart distribution for specific purposes opens way for dynamic calibration. [ABSTRACT FROM AUTHOR]

A series of laboratory experiments are conducted to investigate liquid sloshing in a rectangular tank with a vertical baffle of varying elasticity under horizontal excitations. Image processing techniques are introduced to track baffle displacement, as well as to identify the free surface and the bubble. Dye visualization is employed to reveal the vortex behavior around the baffle. The air entrainment mechanism, including jet impinging, air cavity formation and deformation, and turbulent mixing, is analyzed. Moreover, the study investigates the effect of baffles with different stiffness on suppressing liquid sloshing across a wide range of excitation frequencies. The results show that the peak frequencies of the maximum free surface elevation-frequency curves correspond to the first natural frequency of fluid, the sub-resonance frequency, and the wet natural frequency of the elastic baffle. It is also found that the elastic baffle can effectively reduce liquid sloshing within a certain frequency range and balance the slamming impact both on the wall and the baffle, providing promising prospects for its practical applications. [ABSTRACT FROM AUTHOR]

The utilization of Zernike moments has been extensive in various fields, including image processing and pattern recognition, owing to their desirable characteristics. However, the application of Zernike moments is hindered by two significant obstacles: computational efficiency and accuracy. These issues become particularly noticeable when computing high-order moments. This study presents a novel GPU-based method for efficiently computing high-order Zernike moments by leveraging the computational power of the Single Instruction Multiple Data (SIMD) architecture. The experimental results demonstrate that the proposed method can compute Zernike moments up to order 500 within 0.5 seconds for an image of size 512 × 512. To achieve greater accuracy in Zernike moments computation, a k × k sub-region scheme was incorporated into our approach. The results show that the PSNR value of the Lena image reconstructed from 500-order Zernike moments computed using the 9 × 9 scheme can reach 39.20 dB. • The proposed method has the capability to calculate high-order Zernike Moments. • The proposed method significantly enhances the computational efficiency of Zernike Moments utilizing a GPU. • The proposed method is capable of obtaining accurate Zernike Moments, particularly for high-order moments. [ABSTRACT FROM AUTHOR]

We construct uniform and point-wise asymptotic confidence sets for the single edge in an otherwise smooth image function which are based on rotated differences of two one-sided kernel estimators. Using methods from M-estimation, we show consistency of the estimators of location, slope and height of the edge function and develop a uniform linearization of the contrast process. The uniform confidence bands then rely on a Gaussian approximation of the score process together with anti-concentration results for suprema of Gaussian processes, while point-wise bands are based on asymptotic normality. The finite-sample performance of the point-wise proposed methods is investigated in a simulation study. An illustration to real-world image processing is also given. [ABSTRACT FROM AUTHOR]

Rain effect in images typically is annoying for many multimedia and computer vision tasks. For removing rain effect from a single image, deep leaning techniques have been attracting considerable attentions. This paper designs a novel multi-task leaning architecture in an end-to-end manner to reduce the mapping range from input to output and boost the performance. Concretely, a decomposition net is built to split rain images into clean background and rain layers. Different from previous architectures, our model consists of, besides a component representing the desired clean image, an extra component for the rain layer. During the training phase, we further employ a composition structure to reproduce the input by the separated clean image and rain information for improving the quality of decomposition. Experimental results on both synthetic and real images are conducted to reveal the high-quality recovery by our design, and show its superiority over other state-of-the-art methods. Furthermore, our design is also applicable to other layer decomposition tasks like dust removal. More importantly, our method only requires about 50ms to process a testing image in VGA resolution on a GTX 1080 GPU with promising rain removal quality, making it attractive for practical use. The synthesized dataset and code are publicly available at https://sites.google.com/view/xjguo/rain. [ABSTRACT FROM AUTHOR]

The rate of emulsification in surfactant/oil/water systems is influenced by transport of chemicals and mixing of the fluid phases. In porous media applications, complex flow regimes are generated due to three-dimensional connectivity and irregular cross-sections of the pores facilitating the mixing for emulsification. The properties of the resulting emulsified phase depend on the interplay of flow, mixing and emulsification kinetics of the surfactant/oil/water system. Emulsification can be relatively quick. Direct visualization of the process and compositional gradients in three-dimensional pore space during flow requires imaging at few seconds time intervals. In this study, a flow unit was integrated in a synchrotron beamline-based fast X-ray computed micro-tomography set-up. Non-destructive three-dimensional visualization of multi phase flow inside a porous rock at flow conditions became viable. An oil saturated rock sample was first flooded with water, followed by surfactant solution to mobilize the remaining oil by miscible displacement. The sample was continuously imaged during injection; the scans were made at time intervals of 7–60 s. The presence of an emulsified phase in addition to the oil and the aqueous phases required a more advanced image processing workflow compared to the workflows used for the immiscible fluid systems. A newly developed image processing technique was adopted; the grey levels in the images were correlated with the local oil content in the emulsified fluid regions. The visual extractions of the pore space showed that the emulsification occurred within seconds. Compositional gradients were observed in the emulsified phase as the injected surfactant solution reached the remote locations in the pore space. While a significant fraction of the oil was displaced within few seconds, the compositional gradients persisted over several millimeter length for several minutes, illustrating a sequence of mobilization and solubilization of the oil phase. The ability to interpret such compositional gradients in real time in porous space brings capability to study interfacial phenomena in applications where in situ emulsification occurs under flow. [ABSTRACT FROM AUTHOR]

Radar sensors are considered to be very robust under harsh weather and poor lighting conditions. Largely owing to this reputation, they have found broad application in driver assistance and highly automated driving systems. However, radar sensors have considerably lower precision than cameras. Low sensor precision causes ambiguity in the human interpretation of the measurement data and makes the data labeling process difficult and expensive. On the other hand, without a large amount of high-quality labeled training data, it is difficult, if not impossible, to ensure that the supervised machine learning models can predict, classify, or otherwise analyze the phenomenon of interest with the required accuracy. This paper presents a method for fusing the radar sensor measurements with the camera images. A proposed fully-unsupervised machine learning algorithm converts the radar sensor data to artificial, camera-like, environmental images. Through such data fusion, the algorithm produces more consistent, accurate, and useful information than that provided solely by the radar or the camera. The essential point of the work is the proposal of a novel Conditional Multi-Generator Generative Adversarial Network (CMGGAN) that, being conditioned on the radar sensor measurements, can produce visually appealing images that qualitatively and quantitatively contain all environment features detected by the radar sensor. • Using Generative Adversarial Networks to convert radar measurements to camera images. • Automatic labeling of radar measurements using camera image datasets. • Using radar to increase the camera robustness property. • Using camera to increase the radar precision property. [ABSTRACT FROM AUTHOR]

Abstract Handwritten document transcription aims to obtain the contents of a document to provide efficient information access to, among other, digitised historical documents. The increasing number of historical documents published by libraries and archives makes this an important task. In this context, the use of image processing and understanding techniques in conjunction with assistive technologies reduces the time and human effort required for obtaining the final perfect transcription. The assistive transcription system proposes a hypothesis, usually derived from a recognition process of the handwritten text image. Then, the professional transcriber feedback can be used to obtain an improved hypothesis and speed-up the final transcription. In this framework, a speech signal corresponding to the dictation of the handwritten text can be used as an additional source of information. This multimodal approach, that combines the image of the handwritten text with the speech of the dictation of its contents, could make better the hypotheses (initial and improved) offered to the transcriber. In this paper we study the feasibility of a multimodal interactive transcription system for an assistive paradigm known as Computer Assisted Transcription of Text Images. Different techniques are tested for obtaining the multimodal combination in this framework. The use of the proposed multimodal approach reveals a significant reduction of transcription effort with some multimodal combination techniques, allowing for a faster transcription process. [ABSTRACT FROM AUTHOR]

Abstract In this study, we formulate the concept of "mining maximal-size frequent subgraphs" in the challenging domain of visual data (images and videos). In general, visual knowledge can usually be modeled as attributed relational graphs (ARGs) with local attributes representing local parts and pairwise attributes describing the spatial relationship between parts. Thus, from a practical perspective, such mining of maximal-size subgraphs can be regarded as the discovery of common objects from visual data without given annotations of object bounding boxes. From a theoretical perspective, in this study, we propose a generic definition of common subgraphs among ARGs. Many previous studies can be roughly considered as special cases of the definition. In our definition, we consider 1) variations of unary/pairwise attributes among different ARGs, 2) linkage conditions of different nodes, and 3) the learning of similarity metrics for each node. The generality of our subgraph pattern proposes great challenges to the graph-mining algorithm. We propose an approximate but efficient solution to the mining problem. We conduct five experiments to evaluate our method with different kinds of visual data, including videos and RGB/RGB-D images. These experiments demonstrate the generality of the proposed method. Graphical abstract Highlights • We formulate a generic subgraph pattern for visual data. • Great generality proposes new challenges for graph mining. • We propose an efficient solution to the graph-mining problem. • The generality of our pattern is tested by different visual ARGs. [ABSTRACT FROM AUTHOR]

Abstract Low-light is an inescapable element of our daily surroundings that greatly affects the efficiency of our vision. Research works on low-light imagery have seen a steady growth, particularly in the field of image enhancement, but there is still a lack of a go-to database as a benchmark. Besides, research fields that may assist us in low-light environments, such as object detection, has glossed over this aspect even though breakthroughs-after-breakthroughs had been achieved in recent years, most noticeably from the lack of low-light data (less than 2% of the total images) in successful public benchmark datasets such as PASCAL VOC, ImageNet, and Microsoft COCO. Thus, we propose the Exclusively Dark dataset to elevate this data drought. It consists exclusively of low-light images captured in visible light only, with image and object level annotations. Moreover, we share insightful findings in regards to the effects of low-light on the object detection task by analyzing the visualizations of both hand-crafted and learned features. We found that the effects of low-light reach far deeper into the features than can be solved by simple "illumination invariance". It is our hope that this analysis and the Exclusively Dark dataset can encourage the growth in low-light domain researches on different fields. The dataset can be downloaded at https://github.com/cs-chan/Exclusively-Dark-Image-Dataset. Highlights • A new lowlight image only dataset, the Exclusively DARK is proposed. • Dataset contains 10 lowlight illumination types and annotation of 12 object classes. • Analysis of lowlight images with handcrafted and learned features of object detection. • Lowlight presents illumination challenges that is fundamentally different from bright. • Dataset can be the go-to database to benchmark low-light domain research. [ABSTRACT FROM AUTHOR]

The two-sided quaternion Fourier transform was introduced for the analysis of 2D linear time-invariant partial-differential systems. It has been shown to be a powerful tool in image processing. In this paper, several uncertainty inequalities for the two-sided quaternion Fourier transform are given with optimal constants, including the Pitt's inequality, logarithmic uncertainty inequality, Hausdorff–Young inequality, Hirschman's entropy inequality, generalized Heisenberg inequality, local uncertainty principle and qualitative uncertainty principle. [ABSTRACT FROM AUTHOR]

Abstract Images produced by omnidirectional catadioptric systems provide a larger field of view than conventional cameras. However, these images contain significant radial distortions making classical processing unadapted. In addition, color information is almost neglected in omnidirectional imaging. In this paper, we propose a unifying framework, for central catadioptric color image processing, using Riemannian embedding that deals simultaneously with the geometric deformation due to the use of curved mirrors, and the multi-dimensional characteristic of the image. Based on the introduced Riemannian metric, we derive an adapted Gaussian kernel which is essential in widely used image processing. The resulting new formulation is then applied to various image processing: Image smoothing, Difference of Gaussians filtering and scale-space analysis, edge extraction and corner feature detection using Gaussian derivatives. The experiments illustrate the potential of the proposed approach, and show the higher quality of the adapted processing. Highlights • A new generic metric to process images directly in the catadioptric plane. • The proposed metric encapsulates color information and mirror distorsion. • An adapted Riemannian Gaussian kernel is derived accordingly. • A scale space analysis is performed to compare different metrics. • Effectiveness is proved in smoothing, edge and corner detection. [ABSTRACT FROM AUTHOR]

Coupled effects of hydrodynamics and chemical reaction during gelation process have been studied in the current work. Here, a viscous sodium alginate liquid drop impacts on calcium chloride liquid pool, chemically reacts with the pool, forms a crater, and instantaneously changes its phase from liquid to soft solid called as gel. The drop impingement height and liquid pool concentration are varied to study the effect on this process. This phenomenon is captured in a time resolved high-speed camera and the dynamics of the crater is traced using image processing technique. We also use a mathematical model for crater growth which we assume to be influenced by the rate of gelation. By validating the theoretical trend with the experimental counterpart, the gelation energy has been obtained. Scaling analysis has been executed to determine significant contributory energy in the crater growth. It is also observed that the gel swelling occurs beyond a critical concentration of calcium chloride. Also, the effect of gelation on the homogeneity and strength of alginate gel are interpreted from the surface morphology examination in scanning electron microscope. Thus a new insight of gelation process has been elucidated from the context of fluid dynamics. [ABSTRACT FROM AUTHOR]

Partial nitritation/anammox (PN/A) has been recognized as a cost-efficient process for wastewater nitrogen removal. The addition of carriers could help achieve biomass retention and enhance the treatment efficiency by forming the dense biofilm. However, accurately determining the abundance of anammox bacteria (AnAOB) to evaluate the biofilm development still remains challenging in practice without access to specialized facilities and experimental skills. In this study, we explored the feasibility of utilizing the morphological features of anammox biofilm as an indication of the biofilm development progression, and its correlation with microbial communities was also revealed. The time-series biofilms from an integrated fixed-film activated sludge (IFAS) system with stable PN/A performance were sampled representing the different biofilm development stages. The biofilm morphological features including color and texture were respectively quantified by red (R) coordinate and Local binary pattern (LBP) descriptor via image processing. Hierarchy clustering analysis proved that the extracted morphological descriptors could well distinguish the different stages (colonization, succession, and maturation) of biofilm development. The microbial community dynamics of time-series anammox biofilms were investigated using the amplicon sequence variant (ASV) analysis. Candidatus Brocadia, as the typical AnAOB, dominated in the whole communities of 16.3%–20.0%, moreover, the biofilm development was found to be driven by distinct Brocadia species. Linear regression evidenced that the Brocadia abundance could be directly correlated to the value of R and LBP, and the total variation of microbial communities could be significantly explained by the morphological features via redundancy analysis. This study demonstrates a new way to monitor the biofilm development by extracting the visible features of anammox aggregates, which can help facilitate the automated control of anammox-based bioprocess. [Display omitted] • Biofilm color and texture features were quantified via image processing. • The biofilm development stages could be distinguished by morphological features. • The genus Ca. Brocadia with distinct species drive the biofilm development. • The biofilm morphologic descriptors could correlate with microbial dynamics. [ABSTRACT FROM AUTHOR]

Blind single image super-resolution (SISR) is a challenging task in image processing due to the ill-posed nature of the inverse problem. Complex degradations present in real life images make it difficult to solve this problem using naïve deep learning approaches, where models are often trained on synthetically generated image pairs. Most of the effort so far has been focused on solving the inverse problem under some constraints, such as for a limited space of blur kernels and/or assuming noise-free input images. Yet, there is a gap in the literature to provide a well-generalized deep learning-based solution that performs well on images with unknown and highly complex degradations. In this paper, we propose IKR-Net (Iterative Kernel Reconstruction Network) for blind SISR. In the proposed approach, kernel and noise estimation and high-resolution image reconstruction are carried out iteratively using dedicated deep models. The iterative refinement provides significant improvement in both the reconstructed image and the estimated blur kernel even for noisy inputs. IKR-Net provides a generalized solution that can handle any type of blur and level of noise in the input low-resolution image. IKR-Net achieves state-of-the-art results in blind SISR, especially for noisy images with motion blur. • An iterative deep network architecture is proposed for blind single-image super-resolution. • The model contains separate modules for image reconstruction, blur kernel estimation and noise estimation. • The model achieves state-of-the-art results for noisy images that contain motion blur. [ABSTRACT FROM AUTHOR]

Highlights • Multi-agent system for the classification of people by their age and gender. • The system integrates different techniques for the acquisition and preprocessing of images. • Evaluation of different techniques (Fisherfaces, Eigenfaces, LBP, ANN) and their combination with filters (Gabor, Sobel). • The system operates in environments with uncontrolled conditions (brightness, contrast, saturation or sharpness). Abstract The automatic classification of human images on the basis of age range and gender can be used in audiovisual content adaptation for Smart TVs or marquee advertising. Knowledge about users is used by publishing agencies and departments regulating TV content; on the basis of this information (age, gender) they are able to provide content that suits the interests of users. To this end, the creation of a highly precise image pattern recognition system is necessary, this may be one of the greatest challenges faced by computer technology in the last decades. These recognition systems must apply different pattern recognition techniques, in order to distinct gender and age in the images. In this work, we propose a multi-agent system that integrates different techniques for the acquisition, preprocessing and processing of images for the classification of age and gender. The system has been tested in an office building. Thanks to the use of a multi-agent system which allows to apply different workflows simultaneously, the performance of different methods could be compared (each flow with a different configuration). Experimental results have confirmed that a good preprocessing stage is necessary if we want the classification methods to perform well (Fisherfaces, Eigenfaces, Local Binary Patterns, Multilayer perceptron). The Fisherfaces method has proved to be more effective than MLP and the training time was shorter. In terms of the classification of age, Fisherfaces offers the best results in comparison to the rest of the system's classifiers. The use of filters has allowed to reduce dimensionality, as a result the workload was reduced, a great advantage in a system that performs classification in real time. [ABSTRACT FROM AUTHOR]

Highlights • A survey of exemplar-based texture synthesis methods. • Synthesis approaches are divided into procedural, exemplar-based and model-based methods. • Non-parametric methods generally outperform parametric approaches in the synthesis of regular structured textures. • Parametric and stochastic methods have a better efficiency in the synthesis of irregular textures. Abstract Texture synthesis has become an important topic in image processing, with many fundamental applications in computer vision and image understanding. The purpose of this survey is to give an overview and classification of texture synthesis approaches. According to how they represent, analyze and synthesize textures, the methods are divided into three families; procedural synthesis, exemplar-based synthesis and model-based synthesis, while focusing on exemplar-based methods including most of the synthesis techniques. Finally, experimental evaluations on different textures show that non-parametric synthesis methods lead to the best results when dealing with regular structured textures while parametric algorithms are better for the synthesis of more irregular textures. [ABSTRACT FROM AUTHOR]

In this paper, we explore the different minimal case solutions to the rotational alignment of IMU-camera systems using homography constraints. The assumption that a ground plane is visible in the images can easily be created in many situations. This calibration process is relevant to many smart devices equipped with a camera and an inertial measurement unit (IMU), like micro aerial vehicles (MAVs), smartphones and tablets, and it is a fundamental step for vision and IMU data fusion. Our solutions are novel as they compute the rotational alignment of IMU-camera systems by utilizing a first-order rotation approximation and by solving a polynomial equation system derived from homography constraints. These solutions depend on the calibration case with respect to camera motion (general motion case or pure rotation case) and camera parameters (calibrated camera or partially uncalibrated camera). We then demonstrate that the number of matched points in an image pair can vary from 1.5 to 3. This enables us to calibrate using only one relative movement and provide the exact algebraic solution to the problem. The novel minimal case solutions are useful to reduce the computation time and increase the calibration robustness when using Random Sample Consensus (RANSAC) on the point correspondences between two images. Furthermore, a non-linear parameter optimization over all image pairs is performed. In contrast to the previous calibration methods, our solutions do not require any special hardware, and no problems are experienced with one image pair without special motion. Finally, by evaluating our algorithm on both synthetic and real scene data including data obtained from robots, smartphones and MAVs, we demonstrate that our methods are both efficient and numerically stable for the rotational alignment of IMU-camera systems. [ABSTRACT FROM AUTHOR]

The pointing task is the process of pointing to an object on a computer monitor using a pointing device, or physically touching an object with the hand or finger. It is an important element for users when manipulating visual computer interfaces such as traditional screens and touch-screens. In this context, Fitts’ Law remains one of the central studies that have mathematically modeled the pointing method, and was found to be a good predictor of the average time needed to perform a pointing task. Yet, in our modern computerized society, accessing visual information becomes a central need for all kinds of users, namely users who are blind or visually impaired. Hence, the goal of our study is to evaluate whether Fitts’ Law can be applied for blind candidates using the vibration modality on a touch-screen. To achieve this, we first review the literature on Fitts’ Law and visual data accessibility solutions for blind users. Then, we introduce an experimental framework titled FittsEVAL studying the ability of blind users to tap specific shapes on a touch-screen, while varying different parameters, namely: target distance, target size, and the angle of attack of the pointing task. Experiments on blindfolded and blind candidates show that Fitts’ Law can be effectively applied for blind users using a vibrating touch-screen under certain parameters (i.e., when varying target distance and size), while it is not verified under others (i.e., when varying the angle of attack). This can be considered as a first step toward a more complete experimental evaluation of vibrating touch-screen accessibility, toward designing more adapted interfaces for the blind. [ABSTRACT FROM AUTHOR]

This study considers the 3D human pose estimation problem in a single RGB image by proposing a conditional random field (CRF) model over 2D poses, in which the 3D pose is obtained as a byproduct of the inference process. The unary term of the proposed CRF model is defined based on a powerful heat-map regression network, which has been proposed for 2D human pose estimation. This study also presents a regression network for lifting the 2D pose to 3D pose and proposes the prior term based on the consistency between the estimated 3D pose and the 2D pose. To obtain the approximate solution of the proposed CRF model, the N-best strategy is adopted. The proposed inference algorithm can be viewed as sequential processes of bottom-up generation of 2D and 3D pose proposals from the input 2D image based on deep networks and top-down verification of such proposals by checking their consistencies. To evaluate the proposed method, we use two large-scale datasets: Human3.6M and HumanEva. Experimental results show that the proposed method achieves the state-of-the-art 3D human pose estimation performance. [ABSTRACT FROM AUTHOR]

Projective geometry is a standard mathematical tool for image-based 3D reconstruction. Most reconstruction methods establish pointwise image correspondences using projective geometry. We present an alternative approach based on differential geometry using oriented patches rather than points. Our approach assumes that the scene to be reconstructed is observed by any camera, existing or potential, that satisfies very general conditions, namely, the differentiability of the surface and the bijective projection functions. We show how the notions of the differential geometry such as diffeomorphism, pushforward and pullback are related to the reconstruction problem. A unified theory applicable to various 3D reconstruction problems is presented. Considering two views of the surface, we derive reconstruction equations for oriented patches and pose equations to determine the relative pose of the two cameras. Then we discuss the generalized epipolar geometry and derive the generalized epipolar constraint (compatibility equation) along the epipolar curves. Applying the proposed theory to the projective camera and assuming that affine mapping between small corresponding regions has been estimated, we obtain the minimal pose equation for the case when a fully calibrated camera is moved with its internal parameters unchanged. Equations for the projective epipolar constraints and the fundamental matrix are also derived. Finally, two important nonlinear camera types, the axial and the spherical, are examined. [ABSTRACT FROM AUTHOR]

Estimating the viewpoint of objects in images is an important task for scene understanding. The viewpoint estimation accuracy, however, depends highly on the amount of training data and the quality of the annotation. While humans excel at labelling images with coarse viewpoint annotations like front, back, left or right, the process becomes tedious and the quality of the annotations decreases when finer viewpoint discretisations are required. To solve this problem, we propose a refinement of coarse viewpoint annotations, which are provided by humans, with synthetic data automatically generated from 3D models. To compensate between the difference between synthetic and real images, we introduce a domain adaptation approach that aligns the domain of the synthesized images with the domain of the real images. Experiments show that the proposed approach significantly improves viewpoint estimation on several state-of-the-art datasets. [ABSTRACT FROM AUTHOR]

Plenoptic cameras capture the directional information of the light distribution from a scene. This is accomplished by positioning a microlens array between the main lens and the sensor. This configuration obtains multiple projections for a point in the object space, which allows to retrieve the point’s depth on a single exposure. In recent years, several studies recover depth and shape from the lightfield data using several cues. Nonetheless, references regarding the depth capabilities of a standard plenoptic camera with different zoom and focus settings are scarce. In this work, we formalize a forward projection model and consider projection geometry cues to improve a metric reconstruction methodology for a calibrated standard plenoptic camera. The metric reconstruction methodology is used to evaluate the depth estimation accuracy under certain zoom and focus settings. The reconstruction is applied to new datasets captured for this purpose with objects placed at depths between 0.05 and 2.00 meters. The results indicate that these cameras are able to reconstruct accurately points within the depth range analyzed by appropriately choosing the zoom and focus depth settings. The zoom is a determinant factor on the reconstruction accuracy and the focus depth allows to determine the reconstruction depth range. [ABSTRACT FROM AUTHOR]

The separation of reflection components is an important issue in computer graphics, computer vision, and image processing. This is a highly ill-posed problem since the number of unknowns to solve is much larger than the number of equations. We present a method to reduce the difficulty of this problem by assuming that surface geometry is known. A novel objective function based on robust principal component analysis is proposed to simultaneously separate specularities and estimate the position of light source. We develop an Augmented Lagrangian Multiplier based algorithm to solve the objective function efficiently and effectively. Experimental results on real-world and synthetic data demonstrate the effectiveness of our method. [ABSTRACT FROM AUTHOR]