Filters








5 Hits in 2.0 sec

GridFace: Face Rectification via Learning Local Homography Transformations [article]

Erjin Zhou, Zhimin Cao, Jian Sun
2018 arXiv   pre-print
Our method rectifies the face by local homography transformations, which are estimated by a face rectification network.  ...  We learn the rectification network and recognition network in an end-to-end manner.  ...  Fig. 3 : 3 Local Homography Transformation. The rectification process approximates the 3D face as plane cells and canonicalizes it with local homographies. Fig. 5 : 5 Synthetic 2D Transformations.  ... 
arXiv:1808.06210v1 fatcat:qgfqyig6bbbjrl5cqladwskesi

GridFace: Face Rectification via Learning Local Homography Transformations [chapter]

Erjin Zhou, Zhimin Cao, Jian Sun
2018 Lecture Notes in Computer Science  
Our method rectifies the face by local homography transformations, which are estimated by a face rectification network.  ...  We learn the rectification network and recognition network in an end-to-end manner.  ...  Fig. 3 : 3 Local Homography Transformation. The rectification process approximates the 3D face as plane cells and canonicalizes it with local homographies. Fig. 5 : 5 Synthetic 2D Transformations.  ... 
doi:10.1007/978-3-030-01270-0_1 fatcat:2vagilxd55hbdhxzok5xnyhtdq

Searching for Alignment in Face Recognition [article]

Xiaqing Xu, Qiang Meng, Yunxiao Qin, Jianzhu Guo, Chenxu Zhao, Feng Zhou, Zhen Lei
2021 arXiv   pre-print
Among them, face detection, landmark detection and representation learning have long been studied and a lot of works have been proposed.  ...  A standard pipeline of current face recognition frameworks consists of four individual steps: locating a face with a rough bounding box and several fiducial landmarks, aligning the face image using a pre-defined  ...  (Zhou, Cao, and Sun 2018) use local homography transformations estimated by a rectification network to rectify faces.  ... 
arXiv:2102.05447v2 fatcat:cmqa2fzlmbgxjc327f6oakp5nm

Balanced Alignment for Face Recognition: A Joint Learning Approach [article]

Huawei Wei, Peng Lu, Yichen Wei
2020 arXiv   pre-print
Face alignment is crucial for face recognition and has been widely adopted. However, current practice is too simple and under-explored.  ...  To strike the balance, our second contribution is a novel joint learning approach where alignment learning is controllable with respect to its strength and driven by recognition.  ...  [5] propose a novel non-rigid face rectification method by local homography transformations.  ... 
arXiv:2003.10168v1 fatcat:uo3pxm7klza77koyz74vxumthu

Deep Face Recognition: A Survey [article]

Mei Wang, Weihong Deng
2020 arXiv   pre-print
Since then, deep learning technique, characterized by the hierarchical architecture to stitch together pixels into invariant face representation, has dramatically improved the state-of-the-art performance  ...  Deep learning applies multiple processing layers to learn representations of data with multiple levels of feature extraction.  ...  HOMOGRAPHY TRANSFORMATION OF GRIDFACE [69].  ... 
arXiv:1804.06655v9 fatcat:i2yxh7bf45attlfjd3akyc65me