Fusion between an Algorithm Based on the Characterization of Melanocytic Lesions' Asymmetry with an Ensemble of Convolutional Neural Networks for Melanoma Detection.

Melanoma is still a major health problem worldwide. Early diagnosis is the first step toward reducing its mortality, but it remains a challenge even for experienced dermatologists. Although computer-aided systems have been developed to help diagnosis, the lack of insight into their predictions is still a significant limitation toward acceptance by the medical community. To tackle this issue, we designed handcrafted expert features representing color asymmetry within the lesions, which are parts of the approach used by dermatologists in their daily practice. These features are given to an artificial neural network classifying between nevi and melanoma. We compare our results with an ensemble of 7 state-of-the-art convolutional neural networks and merge the 2 approaches by computing the average prediction. Our experiments are done on a subset of the International Skin Imaging Collaboration 2019 dataset (6296 nevi, 1361 melanomas). The artificial neural network based on asymmetry achieved an area under the curve of 0.873, sensitivity of 90%, and specificity of 67%; the convolutional neural network approach achieved an area under the curve of 0.938, sensitivity of 91%, and specificity of 82%; and the fusion of both approaches achieved an area under the curve of 0.942, sensitivity of 92%, and specificity of 82%. Merging the knowledge of dermatologists with convolutional neural networks showed high performance for melanoma detection, encouraging collaboration between computer science and medical fields.

Encoding the Latent Posterior of Bayesian Neural Networks for Uncertainty Quantification.

Bayesian Neural Networks (BNNs) have long been considered an ideal, yet unscalable solution for improving the robustness and the predictive uncertainty of deep neural networks. While they could capture more accurately the posterior distribution of the network parameters, most BNN approaches are either limited to small networks or rely on constraining assumptions, e.g., parameter independence. These drawbacks have enabled prominence of simple, but computationally heavy approaches such as Deep Ensembles, whose training and testing costs increase linearly with the number of networks. In this work we aim for efficient deep BNNs amenable to complex computer vision architectures, e.g., ResNet-50 DeepLabv3+, and tasks, e.g., semantic segmentation and image classification, with fewer assumptions on the parameters. We achieve this by leveraging variational autoencoders (VAEs) to learn the interaction and the latent distribution of the parameters at each network layer. Our approach, called Latent-Posterior BNN (LP-BNN), is compatible with the recent BatchEnsemble method, leading to highly efficient (in terms of computation and memory during both training and testing) ensembles. LP-BNNs attain competitive results across multiple metrics in several challenging benchmarks for image classification, semantic segmentation, and out-of-distribution detection.

Distinguish self- and hetero-perceived stress through behavioral imaging and physiological features.

Stress reactivity is a complex phenomenon associated to multiple and multimodal expressions. Response to stressors has an obvious survival function and may be seen as an internal regulation to adapt to threat or danger. The intensity of this internal response can be assessed as the self-perception of the stress response. In species with social organization, this response also serves a communicative function, so-called hetero-perception. Our study presents multimodal stress detection assessment - a new methodology combining behavioral imaging and physiological monitoring for analyzing stress from these two perspectives. The system is based on automatic extraction of 39 behavioral (2D+3D video recording) and 62 physiological (Nexus-10 recording) features during a socially evaluated mental arithmetic test. The analysis with machine learning techniques for automatic classification using Support Vector Machine (SVM) show that self-perception and hetero-perception of social stress are both close but different phenomena: self-perception was significantly correlated with hetero-perception but significantly differed from it. Also, assessing stress with SVM through multimodality gave excellent classification results (F1 score values: 0.9±0.012 for hetero-perception and 0.87±0.021 for self-perception). In the best selected feature subsets, we found some common behavioral and physiological features that allow classification of both self- and hetero-perceived stress. However, we also found the contributing features for automatic classifications had opposite distributions: self-perception classification was mainly based on physiological features and hetero-perception was mainly based on behavioral features.

Integration of fuzzy spatial information in tracking based on particle filtering.

In this paper, we propose a novel method to introduce spatial information in particle filters. This information may be expressed as spatial relations (orientation, distance, etc.), velocity, scaling, or shape information. Spatial information is modeled in a generic fuzzy-set framework. The fuzzy models are then introduced in the particle filter and automatically define transition and prior spatial distributions. We also propose an efficient importance distribution to produce relevant particles, which is dedicated to the proposed fuzzy framework. The fuzzy modeling provides flexibility both in the semantics of information and in the transitions from one instant to another one. This allows one to take into account situations where a tracked object changes its direction in a quite abrupt way and where poor prior information on dynamics is available, as demonstrated on synthetic data. As an illustration, two tests on real video sequences are performed in this paper. The first one concerns a classical tracking problem and shows that our approach efficiently tracks objects with complex and unknown dynamics, outperforming classical filtering techniques while using only a small number of particles. In the second experiment, we show the flexibility of our approach for modeling: Fuzzy shapes are modeled in a generic way and allow the tracking of objects with changing shape.

Optical flow computation and visualization in spherical context. Application on 3D+t bio-cellular sequences.

The main focus of this paper is the optical flow computation for 2D movements of objects embedded on 3D surfaces. For spheric-shaped supports, object motion has only two degrees of liberty, thus the 3D optical flow constraint is not relevant. Constancy assumption is formulated using a suitable parametrization of the 3D surface, leading to a 2D equation. Input temporal sequence is also transformed according to the 3D surface parametrization. We build a complete 2D model, taking into account the underlying spherical surface. It has the merit to estimate at a lower cost velocity field in the temporal input sequence. In order to analyze motion computation results, we design an adapted visualization tool, instead of carrying out an inverse transformation for the velocity field. Adapted to the selected parametrization, it displays rapidly moving objects and velocity field and improves the understanding of the displayed information. We display optical flow computation results for 3D+t cell wall simulation sequences.