In this paper, we have proposed a unified pipeline for monocular 3D object detection. By using an anchor-based regression method, we achieved a high-precision dimension and orientation estimation. Then we perform dense sampling in the 3D space and...
In this paper, we have presented a hardness-aware synthesis framework for deep metric learning. Our proposed HDML framework boosts the performance of original metric learning losses by adaptively generating hardness-aware and label-preserving synt...
Credentials & highlights
- Project Views