Multi-Task Feature Decoupling Network with clear division of labor for vehicle component detection

Open Webpage

Yongjie Zhai, Xunqi Zhou, Nianhao Chen, Xuan Liu, Zhenqi Zhang, Xinying Wang, Qianming Wang

Published: 01 Jan 2024, Last Modified: 07 Nov 2025Adv. Eng. Informatics 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•The multi-task feature decoupling network (MTFDN) is proposed.•The method is extendable to other multi-task networks and two-stage algorithms.•Two feature extraction modules are designed according to the task requirements.•A parallel dynamic attention module (PDAM) is proposed.•The detection of vehicle components in complex environment can be realized effectively.