One paper of Robotics and Visual Intelligence (RVI) lab is accepted to IEEE International Conference on Computer Vision (ICCV), 2021, which is one of the top conferences in computer vision!
Existing vanishing point (VP) estimation methods rely on pre-extracted image lines and/or prior knowledge of the number of VPs. However, in practice, this information may be insufficient or unavailable. To solve this problem, we propose a network that treats a perspective image as input and predicts a spherical probability map of VP using the icosahedral spherical representation. Experiments showed that our method achieves the best compromise between generality, accuracy, and efficiency, compared with state-of-the-art approaches.
“Learning Icosahedral Spherical Probability Map Based on Bingham Mixture Model for Vanishing Point Estimation” by Haoang Li*, Kai Chen*, Pyonjin Kim, Kuk-Jin Yoon, Zhe Liu, Kyungdon Joo† and Yun-Hui Liu†
Note: This work is a collaboration with CUHK, Sookmyung Women’s University, KAIST, and University of Cambridge.