Abstract: To address challenges such as large-scale variations, background interference, and occlusions in multi-task autonomous driving perception, this paper proposes YOLOP-MVF, a multi-task ...