Disparity-Based Multiscale Fusion Network for Transportation Detection

The transportation detection of long-distance small objects has low accuracy. In this work, we propose DMF, which is based on disparity depths. We map different disparity regions to 2D candidate regions according to the distance to solve the small-object detection problem. This method clusters disparity maps of different depths. The projected image is extracted with image features in the mapping region. On the one hand, it uses a multicluster method to unsample 2D mapping regions. On the other hand, the feature fusion of different scales is performed on each cluster region. The experimental results on two datasets show that DMF can improve the detection accuracy of small objects.

Automated summary

In this study, a DMF method based on disparity depths is proposed to address the low accuracy issue in transportation detection of long-distance small objects. By mapping different disparity regions to 2D candidate regions, the small-object detection problem is solved. The method clusters disparity maps of different depths and performs feature fusion of different scales. Experimental results on two datasets demonstrate that the DMF method can improve the detection accuracy of small objects.