An investigation of deep learning approaches for efficient assembly component identification

This paper presents a deep learning-based system for automated identification of mechanical fasteners in manufacturing assembly processes. The study compares two object detection algorithms, Mask-RCNN and YOLO v5, for fastener detection using a custom dataset of over 6,000 images. YOLO v5 outperformed Mask-RCNN, achieving 99% mean average precision and demonstrating superior real-time performance, highlighting its potential to enhance automation and reliability in manufacturing operations.

Recommended citation: Ramesh, Kaki, Faisel Mushtaq, Sandip Deshmukh, Tathagata Ray, Chandu Parimi, Ali Basem, and Ammar Elsheikh. "An investigation of deep learning approaches for efficient assembly component identification." Beni-Suef University Journal of Basic and Applied Sciences 13, no. 1 (2024): 79.
Download Paper