Search

This study presents an automated methodology for evaluating micro-channels fabricated using a femtosecond laser on stainless steel substrates. We utilize 3D surface topography and metrological analyses to extract geometric features and detect fabrication defects. Standardized samples were analyzed using a light interferometer, and the resulting data were processed with Principal Component Analysis (PCA) and RANSAC algorithms to derive channel characteristics, such as depth, wall taper, and surface roughness. The proposed method identifies common defects, including bumps and V-defects, which can compromise the functionality of micro-channels. The effectiveness of the approach is validated by comparisons with commercial solutions. This automated procedure aims to enhance the reliability and precision of femtosecond laser micro-milling for industrial applications. The detected defects, combined with fabrication parameters, could be ingested in an AI-based process to optimize fabrication processes. [ABSTRACT FROM AUTHOR]

We present and discuss a runtime architecture that integrates sensorial data and classifiers with a logic-based decision-making system in the context of an e-Health system for the rehabilitation of children with neuromotor disorders. In this application, children perform a rehabilitation task in the form of games. The main aim of the system is to derive a set of parameters the child's current level of cognitive and behavioral performance (e.g., engagement, attention, task accuracy) from the available sensors and classifiers (e.g., eye trackers, motion sensors, emotion recognition techniques) and take decisions accordingly. These decisions are typically aimed at improving the child's performance by triggering appropriate re-engagement stimuli when their attention is low, by changing the game or making it more difficult when the child is losing interest in the task as it is too easy. Alongside state-of-the-art techniques for emotion recognition and head pose estimation, we use a runtime variant of a probabilistic and epistemic logic programming dialect of the Event Calculus, known as the Epistemic Probabilistic Event Calculus. In particular, the probabilistic component of this symbolic framework allows for a natural interface with the machine learning techniques. We overview the architecture and its components, and show some of its characteristics through a discussion of a running example and experiments. [ABSTRACT FROM AUTHOR]

The colorization of grayscale images can, nowadays, take advantage of recent progress and the automation of deep-learning techniques. From the media industry to medical or geospatial applications, image colorization is an attractive and investigated image processing practice, and it is also helpful for revitalizing historical photographs. After exploring some of the existing fully automatic learning methods, the article presents a new neural network architecture, Hyper-U-NET, which combines a U-NET-like architecture and HyperConnections to handle the colorization of historical black and white aerial images. The training dataset (about 10,000 colored aerial image patches) and the realized neural network are available on our GitHub page to boost further research investigations in this field. [ABSTRACT FROM AUTHOR]

Knowledge about the aboveground biomass (AGB) and the diameters at breast height (DBH) distribution can lead to a precise estimation of carbon density and forest structure which can be very important for ecology studies especially for those concerning climate change. In this study, we propose to predict DBH and AGB of individual trees using tree height (H) and crown diameter (CD), and other metrics extracted from airborne laser scanning (ALS) data as input. In the proposed approach, regression methods, such us support vector machine for regression (SVR) and random forests (RF), were used to find a transformation or a transfer function that links the input parameters (H, CD, and other ALS metrics) with the output (DBH and AGB). The developed approach was tested on two datasets collected in southern Norway comprising 3970 and 9467 recorded trees, respectively. The results demonstrate that the developed approach provides better results compared to a state-of-the-art work (based on a linear model with the standard least-squares method) with RMSE equal to 81.4 kg and 92.0 kg, respectively (compared to 94.2 kg and 110.0 kg) for the prediction of AGB, and 5.16 cm and 4.93 cm, respectively (compared to 5.49 cm and 5.30 cm) for DBH. [ABSTRACT FROM AUTHOR]

This paper proposes a novel approach for driving chemometric analyses from spectroscopic data and based on a convolutional neural network (CNN) architecture. For such purpose, the well-known 2-D CNN is adapted to the monodimensional nature of spectroscopic data. In particular, filtering and pooling operations as well as equations for training are revisited. We also propose an alternative to train the resulting 1D-CNN by means of particle swarm optimization. The resulting trained CNN architecture is successively exploited to extract features from a given 1D spectral signature to feed any regression method. In this work, we resorted to 2 advanced and effective methods, which are support vector machine regression and Gaussian process regression. Experimental results conducted on 3 real spectroscopic datasets show the interesting capabilities of the proposed 1D-CNN methods. [ABSTRACT FROM AUTHOR]

The accurate reconstruction of areas obscured by clouds is among the most challenging topics for the remote sensing community since a significant percentage of images archived throughout the world are affected by cloud covers which make them not fully exploitable. The purpose of this paper is to propose new methods to recover missing data in multispectral images due to the presence of clouds by relying on a formulation based on an autoencoder (AE) neural network. We suppose that clouds are opaque and their detection is performed by dedicated algorithms. The AE in our methods aims at modeling the relationship between a given cloud-free image (source image) and a cloud-contaminated image (target image). In particular, two strategies are developed: the first one performs the mapping at a pixel level while the second one at a patch level to take profit from spatial contextual information. Moreover, in order to fix the problem of the hidden layer size, a new solution combining the minimum descriptive length criterion and a Pareto-like selection procedure is introduced. The results of experiments conducted on three different data sets are reported and discussed together with a comparison with reference techniques. [ABSTRACT FROM AUTHOR]

This paper describes three coarse image description strategies, which are meant to promote a rough perception of surrounding objects for visually impaired individuals, with application to indoor spaces. The described algorithms operate on images (grabbed by the user, by means of a chest-mounted camera), and provide in output a list of objects that likely exist in his context across the indoor scene. In this regard, first, different colour, texture, and shape-based feature extractors are generated, followed by a feature learning step by means of AutoEncoder (AE) models. Second, the produced features are fused and fed into a multilabel classifier in order to list the potential objects. The conducted experiments point out that fusing a set of AE-learned features scores higher classification rates with respect to using the features individually. Furthermore, with respect to reference works, our method: (i) yields higher classification accuracies, and (ii) runs (at least four times) faster, which enables a potential full real-time application. [ABSTRACT FROM AUTHOR]

Recently, a new machine learning approach that is termed as the extreme learning machine (ELM) has been introduced in the literature. This approach is characterized by a unified formulation for regression, binary, and multiclass classification problems, and the related solution is given in an analytical compact form. In this letter, we propose an efficient classification method for hyperspectral images based on this machine learning approach. To address the model selection issue that is associated with the ELM, we develop an automatic-solution-based differential evolution (DE). This simple yet powerful evolutionary optimization algorithm uses cross-validation accuracy as a performance indicator for determining the optimal ELM parameters. Experimental results obtained from four benchmark hyperspectral data sets confirm the attractive properties of the proposed DE-ELM method in terms of classification accuracy and computation time. [ABSTRACT FROM PUBLISHER]

This paper proposes a novel approach for the estimation of spectroscopic data by combining the predictions of an ensemble of estimators using the induced ordered weighted averaging (IOWA) fusion operators. For ensemble generation, we use Gaussian process regression (GPR) and extreme learning machine (ELM) estimators associated with different kernels. To render the model selection issue of ELM as efficiently as in the GPR Bayesian estimation method, we develop an automatic solution based on the powerful differential evolution (DE) algorithm. During the fusion process, the IOWA operator needs two things: (1) an order-inducing value; and (2) a way to determine its weights. For the order-inducing value, we propose to use the residual of each estimated output value. Because we cannot compute the true residual, we explore the idea of estimating the residuals themselves by associating to each estimator of the ensemble a second estimator of the same kind called a residual estimator. To learn the weights associated with these nonlinear operators, the proposed method relies on the concept of prioritized aggregation, where we generate the weights directly from the estimated residuals. Experimental results obtained on three real spectroscopic datasets confirm the interesting capabilities of the proposed IOWA fusion method. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

In this letter, we propose to solve the change detection (CD) problem in multitemporal remote-sensing images using interactive segmentation methods. The user needs to input markers related to change and no-change classes in the difference image. Then, the pixels under these markers are used by the support vector machine classifier to generate a spectral-change map. To enhance further the result, we include the spatial contextual information in the decision process using two different solutions based on Markov random field and level-set methods. While the former is a region-driven method, the latter exploits both region and contour for performing the segmentation task. Experiments conducted on a set of four real remote-sensing images acquired by low as well as very high spatial resolution sensors and referring to different kinds of changes confirm the attractive capabilities of the proposed methods in generating accurate CD maps with simple and minimal interaction. [ABSTRACT FROM PUBLISHER]

Light detection and ranging (lidar) data are nowadays a standard data source in studies related to forest ecology and environmental mapping. Medium/high point density lidar data allow to automatically detect individual tree crowns (ITCs), and they provide useful information to predict stem diameter and aboveground biomass of each tree represented by a detected ITC. However, acquisition of field data is necessary for the construction of prediction models that relate field data to lidar data and for validation of such models. When working at ITC level, field data collection is often expensive and time-consuming as accurate tree positions are needed. Active learning (AL) can be very useful in this context as it helps to select the optimal field trees to be measured, reducing the field data collection cost. In this study, we propose a new method of AL for regression based on the minimization of the field data collection cost in terms of distance to navigate between field sample trees, and accuracy in terms of root mean square error of the predictions. The developed method is applied to the prediction of diameter at breast heights (DBH) and aboveground biomass (AGB) of individual trees by using their height and crown diameter as independent variables and support vector regression. The proposed method was tested on two boreal forest datasets, and the obtained results show the effectiveness of the proposed selecting strategy to provide substantial improvements over the different iterations compared to a random selection. The obtained RMSE of DBH/AGB for the first dataset was 5.09 cm/95.5 kg with a cost equal to 8256/6173 m by using the proposed multi-objective method of selection. However, by using a random selection, the RMSE was 5.20 cm/102.1 kg with a cost equal to 28,391/30,086 m. The proposed approach can be efficient in order to get more accurate predictions with smaller costs, especially when a large forest area with no previous field data is subject to inventory and analysis. [ABSTRACT FROM AUTHOR]