Abstract

Abstract
Semantic segmentation consists of classifying each pixel according to a set of classes. Conventional models spend as much effort classifying easy-to-segment pixels as they do classifying hard-to-segment pixels. This is inefficient, especially when deploying to situations with computational constraints. In this work, we propose a framework wherein the model first produces a rough segmentation of the image, and then patches of the image estimated as hard to segment are refined. The framework is evaluated in four datasets (autonomous driving and biomedical), across four state-of-the-art architectures. Our method accelerates inference time by four, with additional gains for training time, at the cost of some output quality.

Abstract
In many real-world prediction tasks, the class labels contain information about the relative order between the labels that are not captured by commonly used loss functions such as multicategory cross-entropy. In ordinal regression, many works have incorporated ordinality into models and loss functions by promoting unimodality of the probability output. However, current approaches are based on heuristics, particularly non-parametric ones, which are still insufficiently explored in the literature. We analyze the set of unimodal distributions in the probability simplex, establishing fundamental properties and giving new perspectives to understand the ordinal regression problem. Two contributions are then proposed to incorporate the preference for unimodal distributions into the predictive model: 1) UnimodalNet, a new architecture that by construction ensures the output is a unimodal distribution, and 2) Wasserstein Regularization, a new loss term that relies on the notion of projection in a set to promote unimodality. Experiments show that the new architecture achieves top performance, while the proposed new loss term is very competitive while maintaining high unimodality.

Abstract
Microscopic techniques in low-to-middle income countries are constrained by the lack of adequate equipment and trained operators. Since light microscopy delivers crucial methods for the diagnosis and screening of numerous diseases, several efforts have been made by the scientific community to develop low-cost devices such as 3D-printed portable microscopes. Nevertheless, these devices present some drawbacks that directly affect image quality: the capture of the samples is done via mobile phones; more affordable lenses are usually used, leading to poorer physical properties and images with lower depth of field; misalignments in the microscopic set-up regarding optical, mechanical, and illumination components are frequent, causing image distortions such as chromatic aberrations. This work investigates several pre-processing methods to tackle the presented issues and proposed a new workflow for low-cost microscopy. Additionally, two new deep learning models based on Convolutional Neural Networks are also proposed (EDoF-CNN-Fast and EDoF-CNN-Pairwise) to generate Extended Depth of Field (EDoF) images, and compared against state-of-the-art approaches. The models were tested using two different datasets of cytology microscopic images: public Cervix93 and a new dataset that has been made publicly available containing images captured with µSmartScope. Experimental results demonstrate that the proposed workflow can achieve state-of-the-art performance when generating EDoF images from low-cost microscopes. © 2022 The Author(s)

Abstract
Robust quantification of animal behavior is fundamental in experimental neuroscience research. Systems providing automated behavioral assessment are an important alternative to manual measurements avoiding problems such as human bias, low reproducibility and high cost. Integrating these tools with closed-loop control systems creates conditions to correlate environment and behavioral expressions effectively, and ultimately explain the neural foundations of behavior. We present an integrated solution for automated behavioral analysis of rodents using deep learning networks on video streams acquired from a depth-sensing camera. The use of depth sensors has notable advantages: tracking/classification performance is improved and independent of animals' coat color, and videos can be recorded in dark conditions without affecting animals' natural behavior. Convolutional and recurrent layers were combined in deep network architectures, and both spatial and temporal representations were successfully learned for a 4-classes behavior classification task (standstill, walking, rearing and grooming). Integration with Arduino microcontrollers creates an easy-to-use control platform providing low-latency feedback signals based on the deep learning automatic classification of animal behavior. The complete system, combining depth-sensor camera, computer, and Arduino microcontroller, allows simple mapping of input-output control signals using the animal's current behavior and position. For example, a feeder can be controlled not by pressing a lever but by the animal behavior itself. An integrated graphical user interface completes a user-friendly and cost-effective solution for animal tracking and behavior classification. This open-software/open-hardware platform can boost the development of customized protocols for automated behavioral research, and support ever more sophisticated, reliable and reproducible behavioral neuroscience experiments.