Comparing modified Yolo V5 and Faster Regional Convolutional Neural Network performance for Recycle Waste Classification

• Published in: 2024 IEEE INTERNATIONAL CONFERENCE ON AUTOMATIC CONTROL AND INTELLIGENT SYSTEMS, I2CACIS 2024
• Main Authors: Hamzah, Raseeda; Ang, Li; Roslan, Rosniza; Teo, Noor Hasimah Ibrahim; Samad, Khairunnisa Abdul; Abu Samah, Khyrina Airin Fariza
• Format: Proceedings Paper
• Language: English
• Published: IEEE 2024
• Subjects: Automation & Control Systems; Computer Science
• Online Access: https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-recordWOS:001308267400074

Waste management research is becoming well-established all over the world. However, there are still improvements needed for developing countries in increasing the effectiveness of waste management. Effective waste management for developing countries is needed to reduce the environmental issues, which have a significant impact. An issue arising from the rise in insect population and diversity of pests is a digestive system problem. Insufficient recycling and waste management practices can have detrimental effects on economic development, resulting in air pollution and health issues. Implementing computer technology such as object recognition has the potential to be advantageous in the field of waste management. Deep learning is now the most widely used approach for object detection. We propose the integration of new modules of Coordinate Attention (CA) mechanism module, K- means++ algorithm and Cascade Shuffle Space to Depth in the Yolo Version 5 to improve the accuracy of the recognition performance. Through the experiments and comparison, the modified version of Yolo v5 perform better performance compared to conventional Yolo V5 and Faster RCNN.