Single-stage detectors for edge computing with CNN architecture to detect animals that suffer the most accidents on Brazilian highways. Using the BRA-Dataset, a specific dataset about these animals, the detectors were trained and validated in the laboratory with edge devices for edge processing for A.I.
This project is a master's project in computer science and computational mathematics at the University of São Paulo (PPG-CCMC, ICMC-USP)
Important pages (Visit the WIKI page)
-
[INSTALL CUDA, OPENCV4 and cuDNN]
-
[YOLOV4 Darknet Model]
-
[YOLOV5 Pytorch Model]
-
[Scaled-YOLOV4]
-
[YOLOR]
-
[YOLOX]
-
[YOLOV7]
For download weights contact the author !
Project Papers:
Understanding the state of the Art in Animal detection and classification using computer vision technologies
DOI: 10.1109/BigData52589.2021.9672049
Brazilian Road’s Animals (BRA): An Image Dataset of Most Commonly Run Over Animals
DOI: 10.1109/SIBGRAPI55357.2022.9991774
Evaluating YOLO architectures for detecting road killed endangered Brazilian animals
DOI: 10.1038/s41598-024-52054-y