People identification using gait information (i.e., the way a person walks) obtained from inertial sensors is robust approach that can be used in multiple situations where vision-based systems are not applicable. Typically, previous methods use hand-crafted features or deep learning approaches with pre-processed as input. In contrast, we present new learning-based end-to-end employs raw data By this way, our able to automatically learn best representations without any constraint introduced...

The task of identifying people by the way they walk is known as `gait recognition'. Although gait mainly used for identification, additional tasks gender recognition or age estimation may be addressed based on well. In such cases, traditional approaches consider those independent ones, defining separated task-specific features and models them. This paper shows that training jointly more than one gait-based tasks, identification converges faster when it trained independently, performance...

The goal of this paper is to identify individuals by analyzing their gait. Instead using binary silhouettes as input data (as done in many previous works) we propose and evaluate the use motion descriptors based on densely sampled short-term trajectories. We take advantage state-of-the-art people detectors define custom spatial configurations around target person, obtaining a rich representation gait motion. local features (described Divergence-Curl-Shear descriptor [M. Jain, H. Jegou P....

Cardiovascular diseases, a leading cause of noncommunicable disease-related deaths, require early and accurate detection to improve patient outcomes. Taking advantage advances in machine learning deep learning, multiple approaches have been proposed the literature address challenge detecting ECG anomalies. Typically, these methods are based on manual interpretation signals, which is time consuming depends expertise healthcare professionals. The objective this work propose system, FADE,...

Gait recognition systems typically rely solely on silhouettes for extracting gait signatures. Nevertheless, these approaches struggle with changes in body shape and dynamic backgrounds; a problem that can be alleviated by learning from multiple modalities. However, many real-life some modalities missing, therefore most existing multimodal frameworks fail to cope missing To tackle this problem, work, we propose UGaitNet, unifying framework recognition, robust UGaitNet handles mingles various...

The goal of this paper is to identify individuals by analyzing their gait. Instead using binary silhouettes as input data (as done in many previous works) we propose and evaluate the use motion descriptors based on densely sampled short-term trajectories. We take advantage state-of-the-art people detectors define custom spatial configurations around target person. Thus, obtaining a pyramidal representation gait motion. local features (described Divergence-Curl-Shear descriptor [1]) extracted...

This work targets people identification in video based on the way they walk (i.e. gait) by using deep learning architectures. We explore use of convolutional neural networks (CNN) for high-level descriptors from low-level motion features optical flow components). The low number training samples each subject and a test set containing subjects different ones makes search good CNN architecture challenging task. carry out thorough experimental evaluation deploying analyzing four distinct models...

Many studies have shown that gait recognition can be used to identify humans at a long distance, with promising results on current datasets. However, those datasets are collected under controlled situations and predefined conditions, which limits the extrapolation of unconstrained in subjects walk freely scenes. To cover this gap, we release novel real-scene dataset (ReSGait), is first scenarios moving not environmental parameters. Overall, our composed 172 870 video sequences, recorded over...

Summary Deep Learning (DL) applications are gaining momentum in the realm of Artificial Intelligence, particularly after GPUs have demonstrated remarkable skills for accelerating their challenging computational requirements. Within this context, Convolutional Neural Network (CNN) models constitute a representative example success on wide set complex applications, datasets where target can be represented through hierarchy local features increasing semantic complexity. In most real scenarios,...

Gait recognition is being employed as an effective approach to identify people without requiring subject collaboration. Nowadays, developed techniques for this task are obtaining high performance on current datasets (usually more than 90 % of accuracy). However, those simple they only contain one in the scene at same time. This fact limits extrapolation results real world conditions where, usually, multiple subjects simultaneously present scene, generating different types occlusions and...

In recent years, biometric systems have positioned themselves among the most widely used technologies for people recognition. this context, palm vein patterns received attention of researchers due to their uniqueness, non-intrusion, and reliability. Currently, research on recognition based deep learning is still very preliminary, works are models by using pre-trained transfer techniques. work, we evaluate end-to-end CNN The proposed method was implemented seven public databases images two...

Palm vein recognition has relevant advantages in comparison with most traditional biometrics, such as high security and performance. In recent years, CNN-based models for vascular biometrics have improved the state-of-the-art, but they disadvantage of requiring a larger number samples training. this context, generation synthetic databases is very effective evaluating performance biometric systems. The present study proposes new perspective transfer learning approach palm recognition, use...