Deep learning

Ideal Binocular Disparity Detectors Learned Using Independent Subspace Analysis on Binocular Natural Image Pairs

An influential theory of mammalian vision, known as the efficient coding hypothesis, holds that early stages in the visual cortex attempts to form an efficient coding of ecologically valid stimuli. Although numerous authors have successfully modelled some aspects of early vision mathematically, closer inspection has found substantial discrepancies between the predictions of some of these models and observations of neurons in the visual cortex. In particular analysis of linear-non-linear models of simple-cells using Independent Component Analysis has found a strong bias towards features on the horoptor. In order to investigate the link between the information content of binocular images, mathematical models of complex cells and physiological recordings, we applied Independent Subspace Analysis to binocular image patches in order to learn a set of complex-cell-like models. We found that these complex-cell-like models exhibited a wide range of binocular disparity-discriminability, although only a minority exhibited high binocular discrimination scores. However, in common with the linear-non-linear model case we found that feature detection was limited to the horoptor suggesting that current mathematical models are limited in their ability to explain the functionality of the visual cortex.

Deep Learning in Label-free Cell Classification

Label-free cell analysis is essential to personalized genomics, cancer diagnostics, and drug development as it avoids adverse effects of staining reagents on cellular viability and cell signaling. However, currently available label-free cell assays mostly rely only on a single feature and lack sufficient differentiation. Also, the sample size analyzed by these assays is limited due to their low throughput. Here, we integrate feature extraction and deep learning with high-throughput quantitative imaging enabled by photonic time stretch, achieving record high accuracy in label-free cell classification. Our system captures quantitative optical phase and intensity images and extracts multiple biophysical features of individual cells. These biophysical measurements form a hyperdimensional feature space in which supervised learning is performed for cell classification. We compare various learning algorithms including artificial neural network, support vector machine, logistic regression, and a novel deep learning pipeline, which adopts global optimization of receiver operating characteristics. As a validation of the enhanced sensitivity and specificity of our system, we show classification of white blood T-cells against colon cancer cells, as well as lipid accumulating algal strains for biofuel production. This system opens up a new path to data-driven phenotypic diagnosis and better understanding of the heterogeneous gene expressions in cells.

Hybrid Network Model for "Deep Learning" of Chemical Data: Application to Antimicrobial Peptides

We present a "deep" network architecture for chemical data analysis and classification together with a prospective proof-of-concept application. The model features a self-organizing map (SOM) as the input layer of a feedforward neural network. The SOM converts molecular descriptors to a two-dimensional image for further processing. We implemented lateral neuron inhibition for contrast enhancement. The model achieved improved classification accuracy and predictive robustness compared to feedforward network classifiers lacking the SOM layer. By nonlinear dimensionality reduction the networks extracted meaningful chemical features from the data and outperformed linear principal component analysis (PCA). The learning machine was trained on the sequence-length independent recognition of antibacterial peptides and correctly predicted the killing activity of a synthetic test peptide against Staphylococcus aureus in an in vitro experiment.

Interactive machine learning for health informatics: when do we need the human-in-the-loop?

Machine learning (ML) is the fastest growing field in computer science, and health informatics is among the greatest challenges. The goal of ML is to develop algorithms which can learn and improve over time and can be used for predictions. Most ML researchers concentrate on automatic machine learning (aML), where great advances have been made, for example, in speech recognition, recommender systems, or autonomous vehicles. Automatic approaches greatly benefit from big data with many training sets. However, in the health domain, sometimes we are confronted with a small number of data sets or rare events, where aML-approaches suffer of insufficient training samples. Here interactive machine learning (iML) may be of help, having its roots in reinforcement learning, preference learning, and active learning. The term iML is not yet well used, so we define it as “algorithms that can interact with agents and can optimize their learning behavior through these interactions, where the agents can also be human.” This “human-in-the-loop” can be beneficial in solving computationally hard problems, e.g., subspace clustering, protein folding, or k-anonymization of health data, where human expertise can help to reduce an exponential search space through heuristic selection of samples. Therefore, what would otherwise be an NP-hard problem, reduces greatly in complexity through the input and the assistance of a human agent involved in the learning phase.

A systematic comparison between visual cues for boundary detection

The detection of object boundaries is a critical first step for many visual processing tasks. Multiple cues (we consider luminance, color, motion and binocular disparity) available in the early visual system may signal object boundaries but little is known about their relative diagnosticity and how to optimally combine them for boundary detection. This study thus aims at understanding how early visual processes inform boundary detection in natural scenes. We collected color binocular video sequences of natural scenes to construct a video database. Each scene was annotated with two full sets of ground-truth contours (one set limited to object boundaries and another set which included all edges). We implemented an integrated computational model of early vision that spans all considered cues, and then assessed their diagnosticity by training machine learning classifiers on individual channels. Color and luminance were found to be most diagnostic while stereo and motion were least. Combining all cues yielded a significant improvement in accuracy beyond that of any cue in isolation. Furthermore, the accuracy of individual cues was found to be a poor predictor of their unique contribution for the combination. This result suggested a complex interaction between cues, which we further quantified using regularization techniques. Our systematic assessment of the accuracy of early vision models for boundary detection together with the resulting annotated video dataset should provide a useful benchmark towards the development of higher-level models of visual processing.

Conceptual and technical advances define a key moment for theoretical neuroscience

Theoretical approaches have long shaped neuroscience, but current needs for theory are elevated and prospects for advancement are bright. Advances in measuring and manipulating neurons demand new models and analyses to guide interpretation. Advances in theoretical neuroscience offer new insights into how signals evolve across areas and new approaches for connecting population activity with behavior. These advances point to a global understanding of brain function based on a hybrid of diverse approaches.

A Mathematical Motivation for Complex-Valued Convolutional Networks

A complex-valued convolutional network (convnet) implements the repeated application of the following composition of three operations, recursively applying the composition to an input vector of nonnegative real numbers: (1) convolution with complex-valued vectors, followed by (2) taking the absolute value of every entry of the resulting vectors, followed by (3) local averaging. For processing real-valued random vectors, complex-valued convnets can be viewed as data-driven multiscale windowed power spectra, data-driven multiscale windowed absolute spectra, data-driven multiwavelet absolute values, or (in their most general configuration) data-driven nonlinear multiwavelet packets. Indeed, complex-valued convnets can calculate multiscale windowed spectra when the convnet filters are windowed complex-valued exponentials. Standard real-valued convnets, using rectified linear units (ReLUs), sigmoidal (e.g., logistic or tanh) nonlinearities, or max pooling, for example, do not obviously exhibit the same exact correspondence with data-driven wavelets (whereas for complex-valued convnets, the correspondence is much more than just a vague analogy). Courtesy of the exact correspondence, the remarkably rich and rigorous body of mathematical analysis for wavelets applies directly to (complex-valued) convnets.

FPGA-Based Stochastic Echo State Networks for Time-Series Forecasting

Hardware implementation of artificial neural networks (ANNs) allows exploiting the inherent parallelism of these systems. Nevertheless, they require a large amount of resources in terms of area and power dissipation. Recently, Reservoir Computing (RC) has arisen as a strategic technique to design recurrent neural networks (RNNs) with simple learning capabilities. In this work, we show a new approach to implement RC systems with digital gates. The proposed method is based on the use of probabilistic computing concepts to reduce the hardware required to implement different arithmetic operations. The result is the development of a highly functional system with low hardware resources. The presented methodology is applied to chaotic time-series forecasting.

A Deep Convolutional Neural Network for segmenting and classifying epithelial and stromal regions in histopathological images

Epithelial (EP) and stromal (ST) are two types of tissues in histological images. Automated segmentation or classification of EP and ST tissues is important when developing computerized system for analyzing the tumor microenvironment. In this paper, a Deep Convolutional Neural Networks (DCNN) based feature learning is presented to automatically segment or classify EP and ST regions from digitized tumor tissue microarrays (TMAs). Current approaches are based on handcraft feature representation, such as color, texture, and Local Binary Patterns (LBP) in classifying two regions. Compared to handcrafted feature based approaches, which involve task dependent representation, DCNN is an end-to-end feature extractor that may be directly learned from the raw pixel intensity value of EP and ST tissues in a data driven fashion. These high-level features contribute to the construction of a supervised classifier for discriminating the two types of tissues. In this work we compare DCNN based models with three handcraft feature extraction based approaches on two different datasets which consist of 157 Hematoxylin and Eosin (H&E) stained images of breast cancer and 1376 immunohistological (IHC) stained images of colorectal cancer, respectively. The DCNN based feature learning approach was shown to have a F1 classification score of 85%, 89%, and 100%, accuracy (ACC) of 84%, 88%, and 100%, and Matthews Correlation Coefficient (MCC) of 86%, 77%, and 100% on two H&E stained (NKI and VGH) and IHC stained data, respectively. Our DNN based approach was shown to outperform three handcraft feature extraction based approaches in terms of the classification of EP and ST regions.

De Novo Design at the Edge of Chaos

Computational medicinal chemistry offers viable strategies for finding, characterizing, and optimizing innovative pharmacologically active compounds. Technological advances in both computer hardware and software as well as biological chemistry have enabled a renaissance of computer-assisted "de novo" design of molecules with desired pharmacological properties. Here, we present our current perspective on the concept of automated molecule generation by highlighting chemocentric methods that may capture druglike chemical space, consider ligand promiscuity for hit and lead finding, and provide fresh ideas for the rational design of customized screening of compound libraries.

Mastering the game of Go with deep neural networks and tree search

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses 'value networks' to evaluate board positions and 'policy networks' to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

Canonical computations of cerebral cortex

The idea that there is a fundamental cortical circuit that performs canonical computations remains compelling though far from proven. Here we review evidence for two canonical operations within sensory cortical areas: a feedforward computation of selectivity; and a recurrent computation of gain in which, given sufficiently strong external input, perhaps from multiple sources, intracortical input largely, but not completely, cancels this external input. This operation leads to many characteristic cortical nonlinearities in integrating multiple stimuli. The cortical computation must combine such local processing with hierarchical processing across areas. We point to important changes in moving from sensory cortex to motor and frontal cortex and the possibility of substantial differences between cortex in rodents vs. species with columnar organization of selectivity.

Kinetic Energy of Hydrocarbons as a Function of Electron Density and Convolutional Neural Networks

We demonstrate a convolutional neural network trained to reproduce the Kohn-Sham kinetic energy of hydrocarbons from an input electron density. The output of the network is used as a nonlocal correction to conventional local and semilocal kinetic functionals. We show that this approximation qualitatively reproduces Kohn-Sham potential energy surfaces when used with conventional exchange correlation functionals. The density which minimizes the total energy given by the functional is examined in detail. We identify several avenues to improve on this exploratory work, by reducing numerical noise and changing the structure of our functional. Finally we examine the features in the density learned by the neural network to anticipate the prospects of generalizing these models.

Expanding perspectives on cognition in humans, animals, and machines

Over the past decade neuroscience has been attacking the problem of cognition with increasing vigor. Yet, what exactly is cognition, beyond a general signifier of anything seemingly complex the brain does? Here, we briefly review attempts to define, describe, explain, build, enhance and experience cognition. We highlight perspectives including psychology, molecular biology, computation, dynamical systems, machine learning, behavior and phenomenology. This survey of the landscape reveals not a clear target for explanation but a pluralistic and evolving scene with diverse opportunities for grounding future research. We argue that rather than getting to the bottom of it, over the next century, by deconstructing and redefining cognition, neuroscience will and should expand rather than merely reduce our concept of the mind.

A combined deep-learning and deformable-model approach to fully automatic segmentation of the left ventricle in cardiac MRI

Segmentation of the left ventricle (LV) from cardiac magnetic resonance imaging (MRI) datasets is an essential step for calculation of clinical indices such as ventricular volume and ejection fraction. In this work, we employ deep learning algorithms combined with deformable models to develop and evaluate a fully automatic LV segmentation tool from short-axis cardiac MRI datasets. The method employs deep learning algorithms to learn the segmentation task from the ground true data. Convolutional networks are employed to automatically detect the LV chamber in MRI dataset. Stacked autoencoders are used to infer the LV shape. The inferred shape is incorporated into deformable models to improve the accuracy and robustness of the segmentation. We validated our method using 45 cardiac MR datasets from the MICCAI 2009 LV segmentation challenge and showed that it outperforms the state-of-the art methods. Excellent agreement with the ground truth was achieved. Validation metrics, percentage of good contours, Dice metric, average perpendicular distance and conformity, were computed as 96.69%, 0.94, 1.81 mm and 0.86, versus those of 79.2-95.62%, 0.87-0.9, 1.76-2.97 mm and 0.67-0.78, obtained by other methods, respectively.

Why neurons mix: high dimensionality for higher cognition

Neurons often respond to diverse combinations of task-relevant variables. This form of mixed selectivity plays an important computational role which is related to the dimensionality of the neural representations: high-dimensional representations with mixed selectivity allow a simple linear readout to generate a huge number of different potential responses. In contrast, neural representations based on highly specialized neurons are low dimensional and they preclude a linear readout from generating several responses that depend on multiple task-relevant variables. Here we review the conceptual and theoretical framework that explains the importance of mixed selectivity and the experimental evidence that recorded neural representations are high-dimensional. We end by discussing the implications for the design of future experiments.

Aerodynamics, sensing and control of insect-scale flapping-wing flight

There are nearly a million known species of flying insects and 13 000 species of flying warm-blooded vertebrates, including mammals, birds and bats. While in flight, their wings not only move forward relative to the air, they also flap up and down, plunge and sweep, so that both lift and thrust can be generated and balanced, accommodate uncertain surrounding environment, with superior flight stability and dynamics with highly varied speeds and missions. As the size of a flyer is reduced, the wing-to-body mass ratio tends to decrease as well. Furthermore, these flyers use integrated system consisting of wings to generate aerodynamic forces, muscles to move the wings, and sensing and control systems to guide and manoeuvre. In this article, recent advances in insect-scale flapping-wing aerodynamics, flexible wing structures, unsteady flight environment, sensing, stability and control are reviewed with perspective offered. In particular, the special features of the low Reynolds number flyers associated with small sizes, thin and light structures, slow flight with comparable wind gust speeds, bioinspired fabrication of wing structures, neuron-based sensing and adaptive control are highlighted.

An Automatic Learning-Based Framework for Robust Nucleus Segmentation

Computer-aided image analysis of histopathology specimens could potentially provide support for early detection and improved characterization of diseases such as brain tumor, pancreatic neuroendocrine tumor (NET), and breast cancer. Automated nucleus segmentation is a prerequisite for various quantitative analyses including automatic morphological feature computation. However, it remains to be a challenging problem due to the complex nature of histopathology images. In this paper, we propose a learning-based framework for robust and automatic nucleus segmentation with shape preservation. Given a nucleus image, it begins with a deep convolutional neural network (CNN) model to generate a probability map, on which an iterative region merging approach is performed for shape initializations. Next, a novel segmentation algorithm is exploited to separate individual nuclei combining a robust selection-based sparse shape model and a local repulsive deformable model. One of the significant benefits of the proposed framework is that it is applicable to different staining histopathology images. Due to the feature learning characteristic of the deep CNN and the high level shape prior modeling, the proposed method is general enough to perform well across multiple scenarios. We have tested the proposed algorithm on three large-scale pathology image datasets using a range of different tissue and stain preparations, and the comparative experiments with recent state of the arts demonstrate the superior performance of the proposed approach.

Neural Information Processing in Cognition: We Start to Understand the Orchestra, but Where is the Conductor?

Research in neural information processing has been successful in the past, providing useful approaches both to practical problems in computer science and to computational models in neuroscience. Recent developments in the area of cognitive neuroscience present new challenges for a computational or theoretical understanding asking for neural information processing models that fulfill criteria or constraints from cognitive psychology, neuroscience and computational efficiency. The most important of these criteria for the evaluation of present and future contributions to this new emerging field are listed at the end of this article.

Deep MRI brain extraction: A 3D convolutional neural network for skull stripping

Brain extraction from magnetic resonance imaging (MRI) is crucial for many neuroimaging workflows. Current methods demonstrate good results on non-enhanced T1-weighted images, but struggle when confronted with other modalities and pathologically altered tissue. In this paper we present a 3D convolutional deep learning architecture to address these shortcomings. In contrast to existing methods, we are not limited to non-enhanced T1w images. When trained appropriately, our approach handles an arbitrary number of modalities including contrast-enhanced scans. Its applicability to MRI data, comprising four channels: non-enhanced and contrast-enhanced T1w, T2w and FLAIR contrasts, is demonstrated on a challenging clinical data set containing brain tumors (N=53), where our approach significantly outperforms six commonly used tools with a mean Dice score of 95.19. Further, the proposed method at least matches state-of-the-art performance as demonstrated on three publicly available data sets: IBSR, LPBA40 and OASIS, totaling N=135 volumes. For the IBSR (96.32) and LPBA40 (96.96) data set the convolutional neuronal network (CNN) obtains the highest average Dice scores, albeit not being significantly different from the second best performing method. For the OASIS data the second best Dice (95.02) results are achieved, with no statistical difference in comparison to the best performing tool. For all data sets the highest average specificity measures are evaluated, whereas the sensitivity displays about average results. Adjusting the cut-off threshold for generating the binary masks from the CNN's probability output can be used to increase the sensitivity of the method. Of course, this comes at the cost of a decreased specificity and has to be decided application specific. Using an optimized GPU implementation predictions can be achieved in less than one minute. The proposed method may prove useful for large-scale studies and clinical trials.

ADAGE-Based Integration of Publicly Available Pseudomonas aeruginosa Gene Expression Data with Denoising Autoencoders Illuminates Microbe-Host Interactions

The increasing number of genome-wide assays of gene expression available from public databases presents opportunities for computational methods that facilitate hypothesis generation and biological interpretation of these data. We present an unsupervised machine learning approach, ADAGE (analysis using denoising autoencoders of gene expression), and apply it to the publicly available gene expression data compendium for Pseudomonas aeruginosa. In this approach, the machine-learned ADAGE model contained 50 nodes which we predicted would correspond to gene expression patterns across the gene expression compendium. While no biological knowledge was used during model construction, cooperonic genes had similar weights across nodes, and genes with similar weights across nodes were significantly more likely to share KEGG pathways. By analyzing newly generated and previously published microarray and transcriptome sequencing data, the ADAGE model identified differences between strains, modeled the cellular response to low oxygen, and predicted the involvement of biological processes based on low-level gene expression differences. ADAGE compared favorably with traditional principal component analysis and independent component analysis approaches in its ability to extract validated patterns, and based on our analyses, we propose that these approaches differ in the types of patterns they preferentially identify. We provide the ADAGE model with analysis of all publicly available P. aeruginosa GeneChip experiments and open source code for use with other species and settings. Extraction of consistent patterns across large-scale collections of genomic data using methods like ADAGE provides the opportunity to identify general principles and biologically important patterns in microbial biology. This approach will be particularly useful in less-well-studied microbial species. IMPORTANCE The quantity and breadth of genome-scale data sets that examine RNA expression in diverse bacterial and eukaryotic species are increasing more rapidly than for curated knowledge. Our ADAGE method integrates such data without requiring gene function, gene pathway, or experiment labeling, making practical its application to any large gene expression compendium. We built a Pseudomonas aeruginosa ADAGE model from a diverse set of publicly available experiments without any prespecified biological knowledge, and this model was accurate and predictive. We provide ADAGE results for the complete P. aeruginosa GeneChip compendium for use by researchers studying P. aeruginosa and source code that facilitates ADAGE's application to other species and data types. Author Video: An author video summary of this article is available.

Cortical Correlates of Low-Level Perception: From Neural Circuits to Percepts

Low-level perception results from neural-based computations, which build a multimodal skeleton of unconscious or self-generated inferences on our environment. This review identifies bottleneck issues concerning the role of early primary sensory cortical areas, mostly in rodent and higher mammals (cats and non-human primates), where perception substrates can be searched at multiple scales of neural integration. We discuss the limitation of purely bottom-up approaches for providing realistic models of early sensory processing and the need for identification of fast adaptive processes, operating within the time of a percept. Future progresses will depend on the careful use of comparative neuroscience (guiding the choices of experimental models and species adapted to the questions under study), on the definition of agreed-upon benchmarks for sensory stimulation, on the simultaneous acquisition of neural data at multiple spatio-temporal scales, and on the in vivo identification of key generic integration and plasticity algorithms validated experimentally and in simulations.

Neural and Computational Mechanisms of Action Processing: Interaction between Visual and Motor Representations

Action recognition has received enormous interest in the field of neuroscience over the last two decades. In spite of this interest, the knowledge in terms of fundamental neural mechanisms that provide constraints for underlying computations remains rather limited. This fact stands in contrast with a wide variety of speculative theories about how action recognition might work. This review focuses on new fundamental electrophysiological results in monkeys, which provide constraints for the detailed underlying computations. In addition, we review models for action recognition and processing that have concrete mathematical implementations, as opposed to conceptual models. We think that only such implemented models can be meaningfully linked quantitatively to physiological data and have a potential to narrow down the many possible computational explanations for action recognition. In addition, only concrete implementations allow judging whether postulated computational concepts have a feasible implementation in terms of realistic neural circuits.

High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning

High-content analysis has revolutionized cancer drug discovery by identifying substances that alter the phenotype of a cell, which prevents tumor growth and metastasis. The high-resolution biofluorescence images from assays allow precise quantitative measures enabling the distinction of small molecules of a host cell from a tumor. In this work, we are particularly interested in the application of deep neural networks (DNNs), a cutting-edge machine learning method, to the classification of compounds in chemical mechanisms of action (MOAs). Compound classification has been performed using image-based profiling methods sometimes combined with feature reduction methods such as principal component analysis or factor analysis. In this article, we map the input features of each cell to a particular MOA class without using any treatment-level profiles or feature reduction methods. To the best of our knowledge, this is the first application of DNN in this domain, leveraging single-cell information. Furthermore, we use deep transfer learning (DTL) to alleviate the intensive and computational demanding effort of searching the huge parameter's space of a DNN. Results show that using this approach, we obtain a 30% speedup and a 2% accuracy improvement.

Neural representation for object recognition in inferotemporal cortex

We suggest that population representation of objects in inferotemporal cortex lie on a continuum between a purely structural, parts-based description and a purely holistic description. The intrinsic dimensionality of object representation is estimated to be around 100, perhaps with lower dimensionalities for object representations more toward the holistic end of the spectrum. Cognitive knowledge in the form of semantic information and task information feed back to inferotemporal cortex from perirhinal and prefrontal cortex respectively, providing high-level multimodal-based expectations that assist in the interpretation of object stimuli. Integration of object information across eye movements may also contribute to object recognition through a process of active vision.

Robust Nucleus/Cell Detection and Segmentation in Digital Pathology and Microscopy Images: A Comprehensive Review

Digital pathology and microscopy image analysis is widely used for comprehensive studies of cell morphology or tissue structure. Manual assessment is labor intensive and prone to interobserver variations. Computer-aided methods, which can significantly improve the objectivity and reproducibility, have attracted a great deal of interest in recent literature. Among the pipeline of building a computer-aided diagnosis system, nucleus or cell detection and segmentation play a very important role to describe the molecular morphological information. In the past few decades, many efforts have been devoted to automated nucleus/cell detection and segmentation. In this review, we provide a comprehensive summary of the recent state-of-the-art nucleus/cell segmentation approaches on different types of microscopy images including bright-field, phase-contrast, differential interference contrast, fluorescence, and electron microscopies. In addition, we discuss the challenges for the current methods and the potential future work of nucleus/cell detection and segmentation.

Using Deep Learning for Image-Based Plant Disease Detection

Crop diseases are a major threat to food security, but their rapid identification remains difficult in many parts of the world due to the lack of the necessary infrastructure. The combination of increasing global smartphone penetration and recent advances in computer vision made possible by deep learning has paved the way for smartphone-assisted disease diagnosis. Using a public dataset of 54,306 images of diseased and healthy plant leaves collected under controlled conditions, we train a deep convolutional neural network to identify 14 crop species and 26 diseases (or absence thereof). The trained model achieves an accuracy of 99.35% on a held-out test set, demonstrating the feasibility of this approach. Overall, the approach of training deep learning models on increasingly large and publicly available image datasets presents a clear path toward smartphone-assisted crop disease diagnosis on a massive global scale.

Evaluating color texture descriptors under large variations of controlled lighting conditions

The recognition of color texture under varying lighting conditions remains an open issue. Several features have been proposed for this purpose, ranging from traditional statistical descriptors to features extracted with neural networks. Still, it is not completely clear under what circumstances a feature performs better than others. In this paper, we report an extensive comparison of old and new texture features, with and without a color normalization step, with a particular focus on how these features are affected by small and large variations in the lighting conditions. The evaluation is performed on a new texture database, which includes 68 samples of raw food acquired under 46 conditions that present single and combined variations of light color, direction, and intensity. The database allows us to systematically investigate the robustness of texture descriptors across large variations of imaging conditions.