Browsing by Autor "Carlos Menacho"

Now showing 1 - 2 of 2

Exploring Edge Computing for Gait Recognition
(2021) Israel Raul Tinini Alvarez; Guillermo Sahonero-Alvarez; Carlos Menacho; Josmar Suarez
Gait Recognition, as a way to identify people, is re-markably attractive for scenarios in which it is not possible to rely on subjects' collaboration. Nevertheless, from all the modalities that Gait Recognition involve, vision-based approaches are better to meet hardware and settings-limitations. Because of that, in the past years, there has been several efforts on developing robust algorithms against visual gait covariates, i.e., view, clothing and carrying variations. However, besides robustness, real-world gait recognition systems also require to be implemented considering near real-time computational demands as well as portability. In this work we propose an Edge Computing approach based on the NVIDIA Jetson Nano development board and the OpenCV OAK-D camera to perform Gait Recognition. To adapt our approach, we created two small data sets that allowed our system to particularize the system to local data. Our pipeline implies the usage of a pre-trained object detection algorithm in the OAK-D, and the execution of both the representation extraction and inference on the Jetson Nano. To test our framework, we first explore its feasibility and consistency in an offline manner. Later, we characterize the complexity and time processing when executing the procedures in an online setup. Our results show that the approach is promising as it allows online operation with an inference time of 35.8 ms.
Fall detection based on CNN models implemented on a mobile robot
(2020) Carlos Menacho; Jhon Ordoñez
Fall accidents are serious events that need to be addressed. Generally, elderly people could suffer these accidents that may lead injures or even death. The use of Convolutional Neural Networks (CNN) has achieved the state of the art for fall detection, but it requires a high computational cost. In this work, we propose an efficient CNN architecture with a reduced number of parameters, which is applied to fall detection in a service with a mobile robot, equipped with a resource-constrained hardware (Nvidia Jetson TX2 platform). Also, different pre-trained CNN models are compared to measure their performances in real scenarios, in addition with other functions like following people and navigation. Furthermore, fall detection is carried out by extraction of temporal features obtained with an Optical Flow extraction from two consecutive RGB images. The proposed network is confirmed by our results to be faster and more suitable for running on resource-constrained Hardware. Our model achieves 88.55% of accuracy using the proposed architecture and it works at 23.16 FPS on GPU and 10.23 FPS on CPU.