no code implementations • 17 Oct 2022 • Shijian Jiang, Guwen Han, Danhang Tang, Yang Zhou, Xiang Li, Jiming Chen, Qi Ye
The decoder aggregate both local image features in pixels and geometric features in vertices.
