From June 7th to 12th, 2015, the 28th IEEE Confference on Computer Vision and Pattern Recognition (CVPR'2015) held in Boston, Massachusetts, United States, Mi Zhang (Supervisor: Professor Yao Jian) from CVRS Lab was invited to participate in this conference and his presentation entitled "Line-Based Multi-Label Energy Optimization for Fisheye Image Rectification and Calibration". After the meeting, Mi Zhang was invited to give a talk in the 18th Doctoral Forum of School of Remote Sensing and Information Engineering, entitled "3D Reconstruction in CVPR 2015", which summarized some technical methods related to the three-dimensional reconstruction of the latest advances in CVPR'2015 meeting, and to share their writing.

The paper of Mi Zhang presenting in CVPR'2015 mainly solved distortion correction and internal camera parameter estimation of large distortion fisheye image. This paper proposed a simultaneous fisheye image distortion rectification and calibration algorithm through a multi-label energy optimization (MLEO) method which can automatically select and extract arcs and can avoid to drop in the local optimum.

CVPR'2015 received a total of 2132 papers, 602 accepted papers, 28.0% acceptance rate. CVPR conference is one of the top international academic conference in the field of computer vision and pattern recognition, and the accepted papers reflect the current maximum level of research in computer vision and pattern recognition.

Live show given by Mi Zhang about his paper.

Talking with Adobe researcher on PatchCut and taking a group photo.

CVPR experience exchange at the Doctor Forum of School of Remote Sensing and Information Engineering

Framework of Fisheye Image Rectification and Calibration