Search

Leveraging inexpensive and human intervention-based annotating methodologies, such as crowdsourcing and web crawling, often leads to datasets with noisy labels. Noisy labels can have a detrimental impact on the performance and generalization of deep neural networks. Robust models that are able to handle and mitigate the effect of these noisy labels are thus essential. In this work, we explore the open challenges of neural network memorization and uncertainty in creating robust learning algorithms with noisy labels. To overcome them, we propose a novel framework called "Bayesian DivideMix++" with two critical components: (i) DivideMix++, to enhance the robustness against memorization and (ii) Monte-Carlo MixMatch, which focuses on improving the effectiveness towards label uncertainty. DivideMix++ improves the pipeline by integrating the warm-up and augmentation pipeline with self-supervised pre-training and dedicated different data augmentations for loss analysis and backpropagation. Monte-Carlo MixMatch leverages uncertainty measurements to mitigate the influence of uncertain samples by reducing their weight in the data augmentation MixMatch step. We validate our proposed pipeline using four datasets encompassing various synthetic and real-world noise settings. We demonstrate the effectiveness and merits of our proposed pipeline using extensive experiments. Bayesian DivideMix++ outperforms the state-of-the-art models by considerable differences in all experiments. Our findings underscore the potential of leveraging these modifications to enhance the performance and generalization of deep neural networks in practical scenarios., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

The vast amount of data produced by today's medical imaging systems has led medical professionals to turn to novel technologies in order to efficiently handle their data and exploit the rich information present in them. In this context, artificial intelligence (AI) is emerging as one of the most prominent solutions, promising to revolutionise every day clinical practice and medical research. The pillar supporting the development of reliable and robust AI algorithms is the appropriate preparation of the medical images to be used by the AI-driven solutions. Here, we provide a comprehensive guide for the necessary steps to prepare medical images prior to developing or applying AI algorithms. The main steps involved in a typical medical image preparation pipeline include: (i) image acquisition at clinical sites, (ii) image de-identification to remove personal information and protect patient privacy, (iii) data curation to control for image and associated information quality, (iv) image storage, and (v) image annotation. There exists a plethora of open access tools to perform each of the aforementioned tasks and are hereby reviewed. Furthermore, we detail medical image repositories covering different organs and diseases. Such repositories are constantly increasing and enriched with the advent of big data. Lastly, we offer directions for future work in this rapidly evolving field., (Copyright © 2021 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.)

Wireless Capsule Endoscopy (WCE) provides a new perspective of the small intestine, since it enables, for the first time, visualization of the entire organ. However, the long visual video analysis time, due to the large number of data in a single WCE study, was an important factor impeding the widespread use of the capsule as a tool for intestinal abnormalities detection. Therefore, the introduction of WCE triggered a new field for the application of computational methods, and in particular, of computer vision. In this paper, we follow the computational approach and come up with a new perspective on the small intestine motility problem. Our approach consists of three steps: first, we review a tool for the visualization of the motility information contained in WCE video; second, we propose algorithms for the characterization of two motility building-blocks: contraction detector and lumen size estimation; finally, we introduce an approach to detect segments of stable motility behavior. Our claims are supported by an evaluation performed with 10 WCE videos, suggesting that our methods ably capture the intestinal motility information., (Copyright © 2015 Elsevier Ltd. All rights reserved.)

Clinical studies on atherosclerosis regression/progression performed by intravascular ultrasound analysis would benefit from accurate alignment of sequences of the same patient before and after clinical interventions and at follow-up. In this article, a methodology for automatic alignment of intravascular ultrasound sequences based on the dynamic time warping technique is proposed. The non-rigid alignment is adapted to the specific task by applying it to multidimensional signals describing the morphologic content of the vessel. Moreover, dynamic time warping is embedded into a framework comprising a strategy to address partial overlapping between acquisitions and a term that regularizes non-physiologic temporal compression/expansion of the sequences. Extensive validation is performed on both synthetic and in vivo data. The proposed method reaches alignment errors of approximately 0.43 mm for pairs of sequences acquired during the same intervention phase and 0.77 mm for pairs of sequences acquired at successive intervention stages., (Copyright © 2013 World Federation for Ultrasound in Medicine & Biology. Published by Elsevier Inc. All rights reserved.)

The Wireless Capsule Endoscopy (WCE) technology allows the visualization of the whole small intestine tract. Since the capsule is freely moving, mainly by the means of peristalsis, the data acquired during the study gives a lot of information about the intestinal motility. However, due to: (1) huge amount of frames, (2) complex intestinal scene appearance and (3) intestinal dynamics that make difficult the visualization of the small intestine physiological phenomena, the analysis of the WCE data requires computer-aided systems to speed up the analysis. In this paper, we propose an efficient algorithm for building a novel representation of the WCE video data, optimal for motility analysis and inspection. The algorithm transforms the 3D video data into 2D longitudinal view by choosing the most informative, from the intestinal motility point of view, part of each frame. This step maximizes the lumen visibility in its longitudinal extension. The task of finding "the best longitudinal view" has been defined as a cost function optimization problem which global minimum is obtained by using Dynamic Programming. Validation on both synthetic data and WCE data shows that the adaptive longitudinal view is a good alternative to the traditional motility analysis done by video analysis. The proposed novel data representation a new, holistic insight into the small intestine motility, allowing to easily define and analyze motility events that are difficult to spot by analyzing WCE video. Moreover, the visual inspection of small intestine motility is 4 times faster then by means of video skimming of the WCE., (Copyright © 2012 Elsevier Ltd. All rights reserved.)

We present a method for automatic struts detection and stent shape estimation in cross-sectional intravascular ultrasound images. A stent shape is first estimated through a comprehensive interpretation of the vessel morphology, performed using a supervised context-aware multi-class classification scheme. Then, the successive strut identification exploits both local appearance and the defined stent shape. The method is tested on 589 images obtained from 80 patients, achieving a F-measure of 74.1% and an averaged distance between manual and automatic struts of 0.10 mm.

We present a fully automatic diagnostic imaging test for Attention-Deficit/Hyperactivity Disorder diagnosis assistance based on previously found evidences of caudate nucleus volumetric abnormalities. The proposed method consists of different steps: a new automatic method for external and internal segmentation of caudate based on Machine Learning methodologies; the definition of a set of new volume relation features, 3D Dissociated Dipoles, used for caudate representation and classification. We separately validate the contributions using real data from a pediatric population and show precise internal caudate segmentation and discrimination power of the diagnostic test, showing significant performance improvements in comparison to other state-of-the-art methods., (Copyright © 2012 Elsevier Ltd. All rights reserved.)

In this paper, we present a fully automatic method which identifies every bifurcation in an intravascular ultrasound (IVUS) sequence, the corresponding frames, the angular orientation with respect to the IVUS acquisition, and the extension. This goal is reached using a two-level classification scheme: first, a classifier is applied to a set of textural features extracted from each image of a sequence. A comparison among three state-of-the-art discriminative classifiers (AdaBoost, random forest, and support vector machine) is performed to identify the most suitable method for the branching detection task. Second, the results are improved by exploiting contextual information using a multiscale stacked sequential learning scheme. The results are then successively refined using a-priori information about branching dimensions and geometry. The proposed approach provides a robust tool for the quick review of pullback sequences, facilitating the evaluation of the lesion at bifurcation sites. The proposed method reaches an F-Measure score of 86.35%, while the F-Measure scores for inter- and intraobserver variability are 71.63% and 76.18%, respectively. The obtained results are positive. Especially, considering the branching detection task is very challenging, due to high variability in bifurcation dimensions and appearance.

Social interactions are a very important component in people's lives. Social network analysis has become a common technique used to model and quantify the properties of social interactions. In this paper, we propose an integrated framework to explore the characteristics of a social network extracted from multimodal dyadic interactions. For our study, we used a set of videos belonging to New York Times' Blogging Heads opinion blog. The Social Network is represented as an oriented graph, whose directed links are determined by the Influence Model. The links' weights are a measure of the "influence" a person has over the other. The states of the Influence Model encode automatically extracted audio/visual features from our videos using state-of-the art algorithms. Our results are reported in terms of accuracy of audio/visual data fusion for speaker segmentation and centrality measures used to characterize the extracted social network.

Vulnerable plaques are the major cause of carotid and coronary vascular problems, such as heart attack or stroke. A correct modeling of plaque echomorphology and composition can help the identification of such lesions. The Rayleigh distribution is widely used to describe (nearly) homogeneous areas in ultrasound images. Since plaques may contain tissues with heterogeneous regions, more complex distributions depending on multiple parameters are usually needed, such as Rice, K or Nakagami distributions. In such cases, the problem formulation becomes more complex, and the optimization procedure to estimate the plaque echomorphology is more difficult. Here, we propose to model the tissue echomorphology by means of a mixture of Rayleigh distributions, known as the Rayleigh mixture model (RMM). The problem formulation is still simple, but its ability to describe complex textural patterns is very powerful. In this paper, we present a method for the automatic estimation of the RMM mixture parameters by means of the expectation maximization algorithm, which aims at characterizing tissue echomorphology in ultrasound (US). The performance of the proposed model is evaluated with a database of in vitro intravascular US cases. We show that the mixture coefficients and Rayleigh parameters explicitly derived from the mixture model are able to accurately describe different plaque types and to significantly improve the characterization performance of an already existing methodology., (© 2011 IEEE)

In this paper, we propose a circular blurred shape model descriptor to deal with the problem of symbol detection and classification as a particular case of object recognition. The feature extraction is performed by capturing the spatial arrangement of significant object characteristics in a correlogram structure. The shape information from objects is shared among correlogram regions, where a prior blurring degree defines the level of distortion allowed in the symbol, making the descriptor tolerant to irregular deformations. Moreover, the descriptor is rotation invariant by definition. We validate the effectiveness of the proposed descriptor in both the multiclass symbol recognition and symbol detection domains. In order to perform the symbol detection, the descriptors are learned using a cascade of classifiers. In the case of multiclass categorization, the new feature space is learned using a set of binary classifiers which are embedded in an error-correcting output code design. The results over four symbol data sets show the significant improvements of the proposed descriptor compared to the state-of-the-art descriptors. In particular, the results are even more significant in those cases where the symbols suffer from elastic deformations.

Speckle noise negatively affects medical ultrasound image shape interpretation and boundary detection. Speckle removal filters are widely used to selectively remove speckle noise without destroying important image features to enhance object boundaries. In this article, a fully automatic bilateral filter tailored to ultrasound images is proposed. The edge preservation property is obtained by embedding noise statistics in the filter framework. Consequently, the filter is able to tackle the multiplicative behavior modulating the smoothing strength with respect to local statistics. The in silico experiments clearly showed that the speckle reducing bilateral filter (SRBF) has superior performances to most of the state of the art filtering methods. The filter is tested on 50 in vivo US images and its influence on a segmentation task is quantified. The results using SRBF filtered data sets show a superior performance to using oriented anisotropic diffusion filtered images. This improvement is due to the adaptive support of SRBF and the embedded noise statistics, yielding a more homogeneous smoothing. SRBF results in a fully automatic, fast and flexible algorithm potentially suitable in wide ranges of speckle noise sizes, for different medical applications (IVUS, B-mode, 3-D matrix array US)., (Copyright 2010 World Federation for Ultrasound in Medicine & Biology. Published by Elsevier Inc. All rights reserved.)

Bioabsorbable drug-eluting coronary stents present a very promising improvement to the common metallic ones solving some of the most important problems of stent implantation: the late restenosis. These stents made of poly-L-lactic acid cause a very subtle acoustic shadow (compared to the metallic ones) making difficult the automatic detection and measurements in images. In this paper, we propose a novel approach based on a cascade of GentleBoost classifiers to detect the stent struts using structural features to code the information of the different subregions of the struts. A stochastic gradient descent method is applied to optimize the overall performance of the detector. Validation results of struts detection are very encouraging with an average F-measure of 81%.

Tissue biomechanical properties (like strain and stress) are playing an increasing role in diagnosis and long-term treatment of intravascular coronary diseases. Their assessment strongly relies on estimation of vessel wall deformation. Since intravascular ultrasound (IVUS) sequences allow visualizing vessel morphology and reflect its dynamics, this technique represents a useful tool for evaluation of tissue mechanical properties. Image misalignment introduced by vessel-catheter motion is a major artifact for a proper tracking of tissue deformation. In this work, we focus on compensating and assessing IVUS rigid in-plane motion due to heart beating. Motion parameters are computed by considering both the vessel geometry and its appearance in the image. Continuum mechanics laws serve to introduce a novel score measuring motion reduction in in vivo sequences. Synthetic experiments validate the proposed score as measure of motion parameters accuracy; whereas results in in vivo pullbacks show the reliability of the presented methodologies in clinical cases.

Three-dimensional intravascular ultrasound (IVUS) allows to visualize and obtain volumetric measurements of coronary lesions through an exploration of the cross sections and longitudinal views of arteries. However, the visualization and subsequent morpho-geometric measurements in IVUS longitudinal cuts are subject to distortion caused by periodic image/vessel motion around the IVUS catheter. Usually, to overcome the image motion artifact ECG-gating and image-gated approaches are proposed, leading to slowing the pullback acquisition or disregarding part of IVUS data. In this paper, we argue that the image motion is due to 3-D vessel geometry as well as cardiac dynamics, and propose a dynamic model based on the tracking of an elliptical vessel approximation to recover the rigid transformation and align IVUS images without loosing any IVUS data. We report an extensive validation with synthetic simulated data and in vivo IVUS sequences of 30 patients achieving an average reduction of the image artifact of 97% in synthetic data and 79% in real-data. Our study shows that IVUS alignment improves longitudinal analysis of the IVUS data and is a necessary step towards accurate reconstruction and volumetric measurements of 3-D IVUS.

The measure of lumen volume on radial arteries can be used to evaluate the vessel response to different vasodilators. In this paper, we present a framework for automatic lumen segmentation in longitudinal cut images of radial artery from Intravascular ultrasound sequences. The segmentation is tackled as a classification problem where the contextual information is exploited by means of Conditional Random Fields (CRFs). A multi-class classification framework is proposed, and inference is achieved by combining binary CRFs according to the Error-Correcting-Output-Code technique. The results are validated against manually segmented sequences. Finally, the method is compared with other state-of-the-art classifiers.

A common way to model multi-class classification problems is by means of Error-Correcting Output Codes (ECOC). Given a multi-class problem, the ECOC technique designs a code word for each class, where each position of the code identifies the membership of the class for a given binary problem. A classification decision is obtained by assigning the label of the class with the closest code. One of the main requirements of the ECOC design is that the base classifier is capable of splitting each sub-group of classes from each binary problem. However, we can not guarantee that a linear classifier model convex regions. Furthermore, non-linear classifiers also fail to manage some type of surfaces. In this paper, we present a novel strategy to model multi-class classification problems using sub-class information in the ECOC framework. Complex problems are solved by splitting the original set of classes into sub-classes, and embedding the binary problems in a problem-dependent ECOC design. Experimental results show that the proposed splitting procedure yields a better performance when the class overlap or the distribution of the training objects conceil the decision boundaries for the base classifier. The results are even more significant when one has a sufficiently large training size.

Despite recovering a normal coronary flow after acute myocardial infarction, percutaneous coronary intervention does not guarantee a proper perfusion (irrigation) of the infarcted area. This damage in microcirculation integrity may detrimentally affect the patient survival. Visual assessment of the myocardium opacification in contrast angiography serves to define a subjective score of the microcirculation integrity myocardial blush analysis (MBA). Although MBA correlates with patient prognosis its visual assessment is a very difficult task that requires of a highly expertise training in order to achieve a good intraobserver and interobserver agreement. In this paper, we provide objective descriptors of the myocardium staining pattern by analyzing the spectrum of the image local statistics. The descriptors proposed discriminate among the different phenomena observed in the angiographic sequence and allow defining an objective score of the myocardial perfusion.

In this paper, we present a general guideline to find a better distance measure for similarity estimation based on statistical analysis of distribution models and distance functions. A new set of distance measures are derived from the harmonic distance, the geometric distance, and their generalized variants according to the Maximum Likelihood theory. These measures can provide a more accurate feature model than the classical Euclidean and Manhattan distances. We also find that the feature elements are often from heterogeneous sources that may have different influence on similarity estimation. Therefore, the assumption of single isotropic distribution model is often inappropriate. To alleviate this problem, we use a boosted distance measure framework that finds multiple distance measures which fit the distribution of selected feature elements best for accurate similarity estimation. The new distance measures for similarity estimation are tested on two applications: stereo matching and motion tracking in video sequences. The performance of boosted distance measure is further evaluated on several benchmark data sets from the UCI repository and two image retrieval applications. In all the experiments, robust results are obtained based on the proposed methods.

Intracoronary UltraSound (IVUS) imaging allows to obtain high resolution images of internal part of coronary arteries. This tool is unique in the possibility to explore internal vessel structures of the coronary wall, being a powerful tool for diagnosis. Since the coronary vessel is moving due to the periodical contraction and expansion of heart muscles, the acquired images present different artifacts. One of the most severe problems is the longitudinal oscillation of the IVUS catheter inside the vessel. To alleviate this problem, ECG-gating has been proposed. The goal of gating is to have subsequent frames that represent the internal vessel section in "stable" position and avoid the repetition of frames; that is to generate an image sequence in which the artifacts due to the heart beat have been removed while, possible translation due to vessel tortuosity can still be present. This paper presents a simple and efficient model of catheter longitudinal movement together with a fast and robust image based gating algorithm. Experimental results on 9 sequences from 7 patients, plus a comparison with ECG gating are presented.

We present a novel approach for retrieval of object categories based on a novel type of image representation: the Generalized Correlogram (GC). In our image representation, the object is described as a constellation of GCs where each one encodes information about some local part and the spatial relations from this part to others (i.e., the part's context). We show how such a representation can be used with fast procedures that learn the object category with weak supervision and efficiently match the model of the object against large collections of images. In the learning stage, we show that by integrating our representation with Boosting the system is able to obtain a compact model that is represented by very few features, where each feature conveys key properties about the object's parts and their spatial arrangement. In the matching step, we propose direct procedures that exploit our representation for efficiently considering spatial coherence between the matching of local parts. Combined with an appropriate data organization such as Inverted Files, we show that thousands of images can be evaluated efficiently. The framework has been applied to different standard databases and we show that our results are favorably compared against state-of-the-art methods in both computational cost and accuracy.

Vessel plaque assessment by analysis of intravascular ultrasound sequences is a useful tool for cardiac disease diagnosis and intervention. Manual detection of luminal (inner) and media-adventitia (external) vessel borders is the main activity of physicians in the process of lumen narrowing (plaque) quantification. Difficult definition of vessel border descriptors, as well as, shades, artifacts, and blurred signal response due to ultrasound physical properties trouble automated adventitia segmentation. In order to efficiently approach such a complex problem, we propose blending advanced anisotropic filtering operators and statistical classification techniques into a vessel border modelling strategy. Our systematic statistical analysis shows that the reported adventitia detection achieves an accuracy in the range of interobserver variability regardless of plaque nature, vessel geometry, and incomplete vessel borders.

We present a heuristic method for learning error correcting output codes matrices based on a hierarchical partition of the class space that maximizes a discriminative criterion. To achieve this goal, the optimal codeword separation is sacrificed in favor of a maximum class discrimination in the partitions. The creation of the hierarchical partition set is performed using a binary tree. As a result, a compact matrix with high discrimination power is obtained. Our method is validated using the UCI database and applied to a real problem, the classification of traffic sign images.

Since the upturn of intravascular ultrasound (IVUS) as an imaging technique for the coronary artery system, much research has been done to simplify the complicated analysis of the resulting images. In this study, an attempt to develop an automatic tissue characterization algorithm for IVUS images was done. The first step was the extraction of texture features. The resulting feature space was used for classification, constructing a likelihood map to represent different coronary plaques. The information in this map was organized using a recently developed geodesic snake formulation, the so-called Stop & Go snake. The novelty of our study lies in this last step, as it was the first time to apply the Stop & Go snake to segment IVUS images.

Among all filtering techniques, those based exclusively on image level sets (geometric flows) have proven to be the less sensitive to the nature of noise and the most contrast preserving. A common feature to existent curvature flows is that they penalize high curvature, regardless of the curve regularity. This constitutes a major drawback since curvature extreme values are standard descriptors of the contour geometry. We argue that an operator designed with shape recovery purposes should include a term penalizing irregularity in the curvature rather than its magnitude. To this purpose, we present a novel geometric flow that includes a function that measures the degree of local irregularity present in the curve. A main advantage is that it achieves non-trivial steady states representing a smooth model of level curves in a noisy image. Performance of our approach is compared to classical filtering techniques in terms of quality in the restored image/shape and asymptotic behavior. We empirically prove that our approach is the technique that achieves the best compromise between image quality and evolution stabilization.

In this work a new statistic deformable model for 3D segmentation of anatomical organs in medical images is proposed. A statistic discriminant snake performs a supervised learning of the object boundary in an image slice to segment the next slice of the image sequence. Each part of the object boundary is projected in a feature space generated by a bank of Gaussian filters. Then, clusters corresponding to different boundary pieces are constructed by means of linear discriminant analysis. Finally, a parametric classifier is generated from each contour in the image slice and embodied into the snake energy-minimization process to guide the snake deformation in the next image slice. The discriminant snake selects and classifies image features by the parametric classifier and deforms to minimize the dissimilarity between the learned and found image features. The new approach is of particular interest for segmenting 3D images with anisotropic spatial resolution, and for tracking temporal image sequences. In particular, several anatomical organs from different imaging modalities are segmented and the results compared to expert tracings.

Purpose: Most published methods directly achieve vessel membrane border detection on cross-sectional intravascular ultrasound (IVUS) images. The vascular structural continuity that exists in entire IVUS image sequences has been overlooked. However, this continuity can have a helpful role in the delineation of vessel membrane contours. To achieve the vessel membrane segmentation more effectively through employing this continuity, a strategy, referred to as multiangle reconstruction, segmentation, and recovery (RSR), is proposed in this paper., Methods: Four main steps are contained in the multiangle-RSR: first, a combination of sampling and interpolation is employed to reconstruct long-axis-model IVUS frames, in which continuity information becomes available. Second, a clustering algorithm is conducted on long-axis-model IVUS frames to roughly extract the media-adventitia (MA) and lumen-intima (LI) boundaries. Third, the segmentation results of cross-sectional IVUS frames are recovered based on the rough results of long-axis-model IVUS frames, and an optimization process that combines downsampling, fitting and smoothing is designed to reduce the interference of bifurcation and side vessels., Results: Multiangle-RSR is tested on a public dataset, and the Hausdorff distance (HD), Jaccard measure (JM), and percentage of area difference (PAD) are utilized as quantitative evaluation metrics. Mean HDs of 0.34 and 0.29 mm are obtained for MA border detection and LI border detection, respectively, which decrease by 43.3% and 9.4%, respectively, compared with their counterparts in previously published approaches. Furthermore, the mean JM is 0.87 for both MA border detection and LI border detection. The mean PADs of the MA contour extraction and the LI contour extraction are 0.10 and 0.11, respectively., Conclusion: The results indicate that the proposed strategy effectively introduces vascular structural continuity by reconstructing long-axis-model IVUS frames and achieves more precise extraction of MA and LI borders., (© 2021 American Association of Physicists in Medicine.)

Abstract: In medical imaging, comparing and retrieving objects is non-trivial because of the high variability in shape and appearance. Such variety leads to poor performance of retrieval algorithms only based on local or global descriptors (shape, color, texture). In this article, we propose a context-based framework for medical image retrieval on the grounds of a global object context based on the mutual positions of local descriptors. This characterization is incorporated into a fast non-rigid registration process to provide invariance against elastic transformations. We apply our method to a complex domain of images—retrieval of intravascular ultrasound images according to vessel morphology. Final results are very encouraging. [Copyright &y& Elsevier]

This paper presents a novel method for improving semantic segmentation performance in computer vision tasks. Our approach utilizes an enhanced UNet architecture that leverages an improved ResNet50 backbone. We replace the last layer of ResNet50 with deformable convolution to enhance feature representation. Additionally, we incorporate an attention mechanism, specifically ECA-ASPP (Attention Spatial Pyramid Pooling), in the encoding path of UNet to capture multi-scale contextual information effectively. In the decoding path of UNet, we explore the use of attention mechanisms after concatenating low-level features with high-level features. Specifically, we investigate two types of attention mechanisms: ECA (Efficient Channel Attention) and LKA (Large Kernel Attention). Our experiments demonstrate that incorporating attention after concatenation improves segmentation accuracy. Furthermore, we compare the performance of ECA and LKA modules in the decoder path. The results indicate that the LKA module outperforms the ECA module. This finding highlights the importance of exploring different attention mechanisms and their impact on segmentation performance. To evaluate the effectiveness of the proposed method, we conduct experiments on benchmark datasets, including Stanford and Cityscapes, as well as the newly introduced WildPASS and DensPASS datasets. Based on our experiments, the proposed method achieved state-of-the-art results including mIoU 85.79 and 82.25 for the Stanford dataset, and the Cityscapes dataset, respectively. The results demonstrate that our proposed method performs well on these datasets, achieving state-of-the-art results with high segmentation accuracy. [ABSTRACT FROM AUTHOR]

• The paper presents an algorithm to fuse multimodality neurological images. • The proposed fusion model is based on an optimized biologically inspired feedforward neural model in two-scale hybrid ℓ 1 − ℓ 0 decomposition domain. • The hyper-parameters of decomposition method and BISFNN model are estimated using GWO technique. • The proposed method provides better-fused images with more detail information compared to the state-of-the-art methods. Medical image fusion plays an important role in the clinical diagnosis of several critical neurological diseases by merging complementary information available in multimodal images. In this paper, a novel CT-MR neurological image fusion framework is proposed using an optimized biologically inspired feedforward neural model in two-scale hybrid ℓ 1 − ℓ 0 decomposition domain using gray wolf optimization to preserve the structural as well as texture information present in source CT and MR images. Initially, the source images are subjected to two-scale ℓ 1 − ℓ 0 decomposition with optimized parameters, giving a scale-1 detail layer, a scale-2 detail layer and a scale-2 base layer. Two detail layers at scale-1 and 2 are fused using an optimized biologically inspired neural model and weighted average scheme based on local energy and modified spatial frequency to maximize the preservation of edges and local textures, respectively, while the scale-2 base layer gets fused using choose max rule to preserve the background information. To optimize the hyper-parameters of hybrid ℓ 1 − ℓ 0 decomposition and biologically inspired neural model, a fitness function is evaluated based on spatial frequency and edge index of the resultant fused image obtained by adding all the fused components. The fusion performance is analyzed by conducting extensive experiments on different CT-MR neurological images. Experimental results indicate that the proposed method provides better-fused images and outperforms the other state-of-the-art fusion methods in both visual and quantitative assessments. [ABSTRACT FROM AUTHOR]

Building facade completion is an important part of digitizing the structures of buildings using computer technology. Due to the intricate textures and structures in building facade images, existing image-completion algorithms cannot accurately restore the rich texture and detailed information. In response, this paper proposes a novel network to simultaneously recover the texture and semantic structural features of building facades. By incorporating dynamic convolutions into each layer of the feature encoder, the shallow layers of the completion network can create a global receptive field, thus enhancing the model's feature-extraction capability. Additionally, a spatial attention branch is integrated into the dynamic convolution module to boost the correlation between the completion area and its surrounding edge area, resulting in improved edge clarity and accuracy of the completed facade image. Experimental results on multiple public image datasets demonstrate that the proposed model in this paper achieves state-of-the-art results when applied to real-world datasets. [ABSTRACT FROM AUTHOR]

A new algorithm is presented for detecting the local weak reflection symmetries in raster images. It uses contours extracted from the segmented image. A convex hull is constructed on the contours, and so-called anchor points are placed on it. The bundles of symmetry line candidates are placed in these points. Each line splits the plane into two open half-planes and arranges the contours into three sets: the first contains the contours pierced by the considered line, while the second and the third include the contours located in one or the other half-plane. The contours are then checked for the reflection symmetry. This means looking for self-symmetries in the first set, and symmetric pairs with one contour in the second set and one contour in the third set. The line which is evaluated as the best symmetry line is selected. After that, the symmetric contours are removed from sets two and three. The remaining contours are then checked again for symmetry. A multi-branch tree representing the hierarchy of the detected local symmetries is the result of the algorithm. [ABSTRACT FROM AUTHOR]

Image-based food pattern classification poses challenges of non-fixed spatial distribution and ingredient occlusion for mainstream computer vision algorithms. However, most current approaches classify food and ingredients by directly extracting abstract features of the entire image through a convolutional neural network (CNN), ignoring the relationship between food and ingredients and ingredient occlusion problem. To address these issues mentioned, we propose a FoodNet for both food and ingredient recognition, which uses a multi-task structure with a multi-scale relationship learning module (MSRL) and a label dependency learning module (LDL). As ingredients normally co-occur in an image, we present the LDL to use the dependency of ingredient to alleviate the occlusion problem of ingredient. MSRL aggregates multi-scale information of food and ingredients, then uses two relational matrixs to model the food-ingredient matching relationship to obtain richer feature representation. The experimental results show that FoodNet can achieve good performance on the Vireo Food-172 and UEC Food-100 datasets. It is worth noting that it reaches the most state-of-the-art level in terms of ingredient recognition in the Vireo Food-172 and UECFood-100.The source code will be made available at https://github.com/visipaper/FoodNet. [ABSTRACT FROM AUTHOR]

Accurately classifying degraded images is a challenging task that relies on domain expertise to devise effective image processing techniques for various levels of degradation. Genetic Programming (GP) has been proven to be an excellent approach for solving image classification tasks. However, the program structures designed in current GP-based methods are not effective in classifying images with quality degradation. During the iterative process of GP algorithms, the high similarity between individuals often results in convergence to local optima, hindering the discovery of the best solutions. Moreover, the varied degrees of image quality degradation often lead to overfitting in the solutions derived by GP. Therefore, this research introduces an innovative program structure, distinct from the traditional program structure, which automates the creation of new features by transmitting information learned across multiple nodes, thus improving GP individual ability in constructing discriminative features. An accompanying evolution strategy addresses high similarity among GP individuals by retaining promising ones, thereby refining the algorithm's development of more effective GP solutions. To counter the potential overfitting issue of the best GP individual, a multi-generational individual ensemble strategy is proposed, focusing on constructing an ensemble GP individual with an enhanced generalization capability. The new method evaluates performance in original, blurry, low contrast, noisy, and occlusion scenarios for six different types of datasets. It compares with a multitude of effective methods. The results show that the new method achieves better classification performance on degraded images compared with the comparative methods. [ABSTRACT FROM AUTHOR]

Glucose transporter-1 (GLUT-1) expression level is a biomarker of tumour hypoxia condition in immunohistochemistry (IHC)-stained images. Thus, the GLUT-1 scoring is a routine procedure currently employed for predicting tumour hypoxia markers in clinical practice. However, visual assessment of GLUT-1 scores is subjective and consequently prone to inter-pathologist variability. Therefore, this study proposes an automated method for assessing GLUT-1 scores in IHC colorectal carcinoma images. For this purpose, we leverage deep transfer learning methodologies for evaluating the performance of six different pre-trained convolutional neural network (CNN) architectures: AlexNet, VGG16, GoogleNet, ResNet50, DenseNet-201 and ShuffleNet. The target CNNs are fine-tuned as classifiers or adapted as feature extractors with support vector machine (SVM) to classify GLUT-1 scores in IHC images. Our experimental results show that the winning model is the trained SVM classifier on the extracted deep features fusion Feat-Concat from DenseNet201, ResNet50 and GoogLeNet extractors. It yields the highest prediction accuracy of 98.86%, thus outperforming the other classifiers on our dataset. We also conclude, from comparing the methodologies, that the off-the-shelf feature extraction is better than the fine-tuning model in terms of time and resources required for training. [ABSTRACT FROM AUTHOR]

If lung cancer is not detected in its initial phases, it can be fatal. However, because of the quantity and structure of its nodules, lung cancer is difficult to detect early. For accurate detections, radiologists require assistance from automated tools. Numerous expert methods have been created over time to assist radiologists in the diagnosis of lung cancer. However, this requires accurate research. Therefore, in this article, we propose a framework to precisely detect lung cancer by categorizing it between benign and malignant nodules. To achieve this objective, an efficient deep-learning algorithm is presented. The presented technique consists of four stages, namely pre-processing, segmentation, classification, and severity stage analysis. Initially, the collected image is given to the pre-processing stage to eliminate the distortion present in the image. Then, the noise-free image is given to the segmentation stage. For segmentation, in this paper, modified regularized K-means (MRKM) clustering algorithm is presented. After the segmentation process, the segmented nodule image is fed to the classification stage to categorize the nodule as benign or malignant (risk nodule). For classification, an improved convolution neural network (ICNN) is presented. The proposed ICNN is designed by modifying CNN with the integration of the adaptive tree seed optimization (ATSO) algorithm. Finally, the stage identification is carried out based on the size of the nodule and we classify the malignant nodule as S1–S4. The presented technique attained the maximum accuracy of 96.5% and performance compared with existing state-of-art methods. [ABSTRACT FROM AUTHOR]

Food and fluid intake monitoring are essential for reducing the risk of dehydration, malnutrition, and obesity. The existing research has been preponderantly focused on dietary monitoring, while fluid intake monitoring, on the other hand, is often neglected. Food and fluid intake monitoring can be based on wearable sensors, environmental sensors, smart containers, and the collaborative use of multiple sensors. Vision-based intake monitoring methods have been widely exploited with the development of visual devices and computer vision algorithms. Vision-based methods provide non-intrusive solutions for monitoring. They have shown promising performance in food/beverage recognition and segmentation, human intake action detection and classification, and food volume/fluid amount estimation. However, occlusion, privacy, computational efficiency, and practicality pose significant challenges. This paper reviews the existing work (253 articles) on vision-based intake (food and fluid) monitoring methods to assess the size and scope of the available literature and identify the current challenges and research gaps. This paper uses tables and graphs to depict the patterns of device selection, viewing angle, tasks, algorithms, experimental settings, and performance of the existing monitoring systems. [ABSTRACT FROM AUTHOR]

Tracking points in ultrasound (US) videos can be especially useful to characterize tissues in motion. Tracking algorithms that analyze successive video frames, such as variations of Optical Flow and Lucas–Kanade (LK), exploit frame-to-frame temporal information to track regions of interest. In contrast, convolutional neural-network (CNN) models process each video frame independently of neighboring frames. In this paper, we show that frame-to-frame trackers accumulate error over time. We propose three interpolation-like methods to combat error accumulation and show that all three methods reduce tracking errors in frame-to-frame trackers. On the neural-network end, we show that a CNN-based tracker, DeepLabCut (DLC), outperforms all four frame-to-frame trackers when tracking tissues in motion. DLC is more accurate than the frame-to-frame trackers and less sensitive to variations in types of tissue movement. The only caveat found with DLC comes from its non-temporal tracking strategy, leading to jitter between consecutive frames. Overall, when tracking points in videos of moving tissue, we recommend using DLC when prioritizing accuracy and robustness across movements in videos, and using LK with the proposed error-correction methods for small movements when tracking jitter is unacceptable. [ABSTRACT FROM AUTHOR]

Feature selection is an important task in the areas of pattern recognition and data mining. Various approaches to feature selection have been developed. In particular, this paper focuses on the algorithms for computing irreducible testors, which have been used to solve feature selection problems. The calculation of irreducible testors is an expensive computational process; the complexity of the algorithms to calculate the complete set of irreducible testors exponentially depends on the number of characteristics that describe the objects in the problem. To improve the execution time of these algorithms, different alternatives have been developed, such as parallel implementations, hardware-software implementation, rearrangement of the data, as well as heuristics to generate just an irreducible testor or a subset of the entire set of irreducible testors, among other strategies. This paper presents a review of the literature on irreducible testors, with the aim of providing a guide for researchers working in the areas of pattern recognition and data mining, interested in feature selection, using heterogeneous data and possibly missing data. [ABSTRACT FROM AUTHOR]

We have previously developed an original method to evaluate small bowel motor function based on computer vision analysis of endoluminal images obtained by capsule endoscopy. Our aim was to demonstrate intestinal motor abnormalities in patients with functional bowel disorders by endoluminal vision analysis. Patients with functional bowel disorders (n = 205) and healthy subjects (n = 136) ingested the endoscopic capsule (Pillcam-SB2, Given-Imaging) after overnight fast and 45 min after gastric exit of the capsule a liquid meal (300 ml, 1 kcal/ml) was administered. Endoluminal image analysis was performed by computer vision and machine learning techniques to define the normal range and to identify clusters of abnormal function. After training the algorithm, we used 196 patients and 48 healthy subjects, completely naive, as test set. In the test set, 51 patients (26%) were detected outside the normal range (P < 0.001 vs. 3 healthy subjects) and clustered into hypo- and hyperdynamic subgroups compared with healthy subjects. Patients with hypodynamic behavior (n = 38) exhibited less luminal closure sequences (41 ± 2% of the recording time vs. 61 ± 2%; P < 0.001) and more static sequences (38 ± 3 vs. 20 ± 2%; P < 0.001); in contrast, patients with hyperdynamic behavior (n = 13) had an increased proportion of luminal closure sequences (73 ± 4 vs. 61 ± 2%; P = 0.029) and more high-motion sequences (3 ± 1 vs. 0.5 ± 0.1%; P < 0.001). Applying an original methodology, we have developed a novel classification of functional gut disorders based on objective, physiological criteria of small bowel function. [ABSTRACT FROM AUTHOR]

With the rapid development of intelligent transportation, more and more vehicles are equipped with intelligent traffic sign recognition systems, which can reduce the potential safety hazards caused by human cognitive errors. Therefore, a more safe and reliable traffic sign recognition system is the demand of drivers, and it is also the research hotspot of current automobile manufacturers. However, the pictures taken by the actual driving car are inevitably distorted and blurred. In addition, there are external uncontrollable factors, such as the impact of bad weather, which make the research of traffic sign recognition system face many difficulties, and the practical application is far from mature. In order to solve the above challenges, this paper proposes a Yolo model for traffic sign recognition. Firstly, the traffic signs are roughly divided into several categories and then preprocessed according to the characteristics of various types of signs. The processed pictures are input into the optimized convolutional neural network to subdivide the categories to obtain the specific categories. Finally, the proposed recognition algorithm is tested with the data set based on the German traffic sign recognition standard and compared with other baseline algorithms. The results show that the algorithm greatly improves the running speed on the basis of ensuring a high classification accuracy and is more suitable for traffic sign recognition system. [ABSTRACT FROM AUTHOR]

Many Error-Correcting Output Codes (ECOC) algorithms had been proposed based on the hard coding (HC) schemes: binary coding {1, 0} or ternary coding {+ 1, -1, 0}. This paper introduces two novel strategies to recode the original code matrices with the mean values and the intervals of learners' outputs, which are named Mean Value Recoding (MVR) and Interval Recoding (IR) strategies. Both strategies are designed to reduce the distance between the outputs of base learners and the target codewords, aiming to produce more accurate results compared with the HC schemes. It is the first time that two concepts, soft recoding, and learner dependent, are injected into the ECOC framework to the best of our knowledge. To verify the effectiveness of our strategies, four data-independent ECOC algorithms and two data-dependent ECOC algorithms are deployed in the experiments based on UCI data sets. The experiments are carried out using the original HC strategies and our soft recoding strategies, and results verify that our strategies outperform the HC-based algorithms in most cases by producing balanced results among classes. In short, our strategies can improve the performance of different ECOC algorithms. Our python code and the corresponding data sets are available for non-commercial or research use at: https://github.com/MLDMXM2017/softcoding-ECOC. [ABSTRACT FROM AUTHOR]

To optimally improve signal-to-noise ratio in noisy environments, a hearing assistance device must correctly identify what is signal and what is noise. Many of the biosignal-based approaches to solving this question are themselves subject to noise, but head angle is an overt behavior that may be possible to capture in practical devices in the real world. Previous orientation studies have demonstrated that head angle is systematically related to listening target; our study aimed to examine whether this relationship is sufficiently reliable to be used in group conversations where participants may be seated in different layouts and the listener is free to turn their body as well as their head. In addition to this simple method, we developed a source-selection algorithm based on a hidden Markov model (HMM) trained on listeners' head movement. The performance of this model and the simple head-steering method was evaluated using publicly available behavioral data. Head angle during group conversation was predictive of active talker, exhibiting an undershoot with a slope consistent with that found in simple orientation studies, but the intercept of the linear relationship was different for different talker layouts, suggesting it would be problematic to rely exclusively on this information to predict the location of auditory attention. Provided the location of all target talkers is known, the HMM source selection model implemented here, however, showed significantly lower error in identifying listeners' auditory attention than the linear head-steering method. [ABSTRACT FROM AUTHOR]

Error-Correcting Output Codes has become a well-known, established technique for multiclass classification due to its simplicity and efficiency. Each binary split contains different original classes. A noncompetent classifier emerges when it classifies an instance whose real class does not belong to the metasubclasses which is used to learn the classifier. How to reduce the error caused by the noncompetent classifiers under diversity big enough is urgent for ECOC classification. The weighted decoding strategy can be used to reduce the error caused by the noncompetence contradiction through relearning the weight coefficient matrix. To this end, a new weighted decoding strategy taking the classifier competence reliability into consideration is presented in this paper, which is suitable for any coding matrix. Support Vector Data Description is applied to compute the distance from an instance to the metasubclasses. The distance reflects the competence reliability and is fused as the weight in the base classifier combination. In so doing, the effect of the competent classifiers on classification is reinforced, while the bias induced by the noncompetent ones is decreased. Reflecting the competence reliability, the weights of classifiers for each instance change dynamically, which accords with the classification practice. The statistical simulations based on benchmark datasets indicate that our proposed algorithm outperforms other methods and provides new thought for solving the noncompetence problem. [ABSTRACT FROM AUTHOR]

Skin cancers are increasing at an alarming rate, and detection in the early stages is essential for advanced treatment. The current segmentation methods have limited labeling ability to the ground truth images due to the numerous noisy expert annotations present in the datasets. The precise boundary segmentation is essential to correctly locate and diagnose the various skin lesions. In this work, the lesion segmentation method is proposed as a Markov decision process. It is solved by training an agent to segment the region using a deep reinforcement-learning algorithm. Our method is similar to the delineation of a region of interest by the physicians. The agent follows a set of serial actions for the region delineation, and the action space is defined as a set of continuous action parameters. The segmentation model learns in continuous action space using the deep deterministic policy gradient algorithm. The proposed method enables continuous improvement in performance as we proceed from coarse segmentation results to finer results. Finally, our proposed model is evaluated on the International Skin Imaging Collaboration (ISIC) 2017 image dataset, Human against Machine (HAM10000), and PH2 dataset. On the ISIC 2017 dataset, the algorithm achieves an accuracy of 96.33% for the naevus cases, 95.39% for the melanoma cases, and 94.27% for the seborrheic keratosis cases. The other metrics are evaluated on these datasets and rank higher when compared with the current state-of-the-art lesion segmentation algorithms. [ABSTRACT FROM AUTHOR]

In comparison with static image object detection, focusing on video objects has greater research significance in realizing intelligent monitoring and automatic anomaly detection. However, it may be challenging to apply the most advanced image recognition networks to video data, as the number of static frame files represented in a video is often huge, thereby causing the problem of the slow evaluation speed, in addition to other issues, such as motion blur, low resolution, occlusion, and object deformation. In the present study, to mitigate these deficiencies, we applied sparse feature propagation to improve the detection speed and dense feature aggregation to refine the detection accuracy. Moreover, we utilized the key frame scheduling strategy relying on the consistency of feature information. Implementing these technologies allowed steadily improving the detection speed and accuracy to achieve high performance. To verify the applicability of the optimized video detection strategy proposed in this paper, we selected the part of the video data in the ImageNet VID training dataset. Then, the other part of this dataset was used to conduct the experiments, including the calculation and comparison of mean average precision (MAP) and frames per second (FPS). [ABSTRACT FROM AUTHOR]

Robotic exoskeletons are developed with the aim of enhancing convenience and physical possibilities in daily life. However, at present, these devices lack sufficient synchronization with human movements. To optimize human-exoskeleton interaction, this article proposes a gait recognition and prediction model, called the gait neural network (GNN), which is based on the temporal convolutional network. It consists of an intermediate network, a target network, and a recognition and prediction model. The novel structure of the algorithm can make full use of the historical information from sensors. The performance of the GNN is evaluated based on the publicly available HuGaDB dataset, as well as on data collected by an inertial-based wearable motion capture device. The results show that the proposed approach is highly effective and achieves superior performance compared with existing methods. [ABSTRACT FROM AUTHOR]

When training base classifier by ternary Error Correcting Output Codes (ECOC), it is well know that some classes are ignored. On this account, a non-competent classifier emerges when it classify an instance whose real label does not belong to the meta-subclasses. Meanwhile, the classic ECOC dichotomizers can only produce binary outputs and have no capability of rejection for classification. To overcome the non-competence problem and better model the multi-class problem for reducing the classification cost, we embed reject option to ECOC and present a new variant of ECOC algorithm called as Reject-Option-based Re-encoding ECOC (ROECOC). The cost-sensitive classification model and cost-loss function based on Receiver Operating Characteristic (ROC) curve are built respectively. The optimal reject threshold values are obtained by combing the condition to be met for minimizing the loss function and the ROC convex hull. In so doing, reject option (t1, t2) provides a three-symbol output to make dichotomizers more competent and ROECOC more universal and practical for cost-sensitive classification issue. Experimental results on two kinds of datasets show that our scheme with low-degree freedom of initialized ECOC can effectively enhance accuracy and reduce cost. [ABSTRACT FROM AUTHOR]

When the English teaching text is regarded as the ontology, it must involve how to describe the attribute effectively. However, in the current research, the research on the automatic extraction of labels for English teaching texts is still insufficient. Intelligent English teaching has become an inevitable trend in the development of future English teaching models, so it is necessary to cooperate with intelligent text recognition technology. Based on SVM, this study applies convolutional neural network algorithm to text recognition of English teaching content, and effectively recognizes text features. After feature extraction, the original text content has been changed into data that the machine can directly identify and analyze, and semantic analysis is performed. In order to verify the performance of the algorithm, the performance of the algorithm was analyzed by example verification. It can be seen from the results that the proposed method has a certain accuracy rate and can be applied to the text recognition classification of English teaching content and can provide reference direction for related research. [ABSTRACT FROM AUTHOR]

Online education has become an important way of learning English at present, and English vocabulary teaching can improve the efficiency of English vocabulary teaching through target visual detection. However, from the existing research, it can be seen that there are still some shortcomings in English vocabulary recognition. In order to improve the English vocabulary recognition effect, based on machine learning recognition technology, this study combines English vocabulary recognition needs of online education to construct an English vocabulary detection model based on convolutional neural network. The model takes the word's overall feature as the feature extraction principle and adopts the analysis and extraction of the joint segment feature. Moreover, it discards the complicated process of first dividing a single letter and then performing feature extraction and recognition. In addition, this study design example tests to perform algorithm performance analysis. The experimental results show that the proposed algorithm model has certain effects, and it can be used as an auxiliary algorithm for online English vocabulary teaching. [ABSTRACT FROM AUTHOR]