Pedestrian detection method and Research of Attention Mechanism in Traffic Scene
Fanshen Xu · Highlights in Science Engineering and Technology · 2022
Aiming at the occlusion problem when pedestrian detection algorithms are applied in traffic scenes, an occlusion-aware pedestrian 2D saliency texture operator TS-LBP detection algorithm combined with dual attention mechanism is proposed. The algorithm first uses the differential feature-aware fusion method to mine the complementary information between multi-modal features to optimize the channel features; then uses the detection framework with an efficient anchor-free mechanism to greatly reduce the computational complexity of the model and improve the detection speed. The experimental results show that the method in this paper is fast and accurate, and has a good detection effect.