Electrotechnical and Computer Engineering
Vol. 38 No. 07 (2023): Proceedings of Faculty of Technical Sciences
DETECTION AND SEGMENTATION OF KAYAKERS USING CONVOLUTIONAL NEURAL NETWORKS
Abstract
The paper presents a system for detecting kayakers on video. The system parses the video and processes each frame. In the images, instances of kayakers are detected and segmented. To solve the mentioned tasks, the Mask R-CNN method is used with ResNet101 architecture. The model was created using the transfer learning technique. The transfer learning technique uses a Mask R-CNN model previously trained on the Microsoft COCO dataset. As a result of the system, an output video ws generated with detected and segmented kayaker.
References
[1] Abdulla, W.: Mask R-CNN for object detection and instance segmentation on keras and tensorflow. https://github.com/matterport/Mask RCNN (2017), accessed 07-Oct-2020
[2] T. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll ́ar, and C. L. Zitnick, “Microsoft COCO: Common objects in context,” in ECCV, 2014.
[3] Girshick, Ross & Donahue, Jeff & Darrell, Trevor & Malik, Jitendra. (2015). Region-Based Convolutional Networks for Accurate Object Detection and Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 38. 1-1. 10.1109/TPAMI.2015.2437384.
[4] Handalage, Upulie & Kuganandamurthy, Lakshini. (2021). Real-Time Object Detection Using YOLO: A Review. 10.13140/RG.2.2.24367.66723.
[5] VGG Image Annotator (VIA) https://www.robots.ox.ac.uk/~vgg/software/via/via_demo.html [приступљено 24.11.2022.]
[6] Matterport Inc. Mask R-CNN https://github.com/matterport/Mask_RCNN [приступљено 24.11.2022.]