The Nuts and Bolts of Intelligent Sorting: An Improved Dynamic Sorting Algorithm Based on ResNet101
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Intelligent sorting of industrial components remains challenging due to object similarity,mutual occlusion, and dynamic motion. To address these issues, this paper proposes an improved object detection method for industrial sorting based on ResNet101. A multidimensional attention mechanism is introduced to enhance foreground features, while spatial pyramid pooling and global average pooling are incorporated to improve robustness. An optimized loss function with modulation parameters accelerates convergence. Experiments show a recognition accuracy of 98%, an average detection time of 231.21 ms, and a dynamic grasping success rate of 86.5\%.
Fields of Study & Subjects
Advanced Neural Network Applications
Industrial Vision Systems and Defect Detection
Robot Manipulation and Learning
Publication details
- Journal / source
- International Conference on Electrical Automation and Artificial Intelligence (ICEAAI)
- Pages
- 1158–1161
- Publication date
- 16 January 2026
- Language
- EN