Abstract:Face alignment and recognition in less controlled environment are one of the most essential bottlenecks for practical face recognition system. Recently several researches have focused on partial face recognition problem, but few works have addressed the problem of face alignment under partial occlusion. In this paper, we present a robust face alignment method by combining local feature matching and Probabilistic Hough Transform (PHT) for partial face alignment in near infrared (NIR) images. Given a set of well aligned faces as target, and for face images with occlusions, their correspondences are established by local feature matching. For faces with missing components, many false matches of local features will be built due to lack of holistic information. The PHT approach aims to find correct correspondences and resist the inevitable false ones by taking each parameter candidate generated by correspondences pair as a vote in the 4-D in-plane transform parameter space. We also employ geometric constraints and appearance consistency and combine them with PHT in an probabilistic hough optimization function, so that each vote is weighted by a probabilistic score. Experiments of alignment on both MBGC portal face video and facial images with Glass-face occlusions show that our approach can reliably and accurately deal with missing data of facial components caused by partial occlusion.

Recognition of Partially Occluded Polygon from Line Drawing

Face recognition with contiguous occlusion using linear regression and level set method

A New Method of Recognizing Polygons under Perspective Transformation

Partially Occluded Object Recognition Algorithm Based on Feature Description Integrity

Local Feature-Based Recognition of Partially Occluded Objects Using Neural Network

Partially occluded object matching via multi-level description and evaluation of contour features

The recognition of 2D partially occluded objects using curve moment invariants

Face Recognition with Partial Occlusion.

Recognition of Polygon Based on Homograph Principle

Facial Expression Recognition under Partial Occlusion Based on Fusion of Global and Local Features

Face Recognition under Occlusions and Variant Expressions with Partial Similarity

An efficient method for occluded face recognition

Partial Occlusion Robust Object Tracking Using an Effective Appearance Model

Improved Approximate Algorithm of Polygon

Face Alignment under Partial Occlusion in near Infrared Images

A New Method for Polygon Detection Based on Clustering

Aligning Partially Overlapping Point Sets: an Inner Approximation Algorithm

Topology Preserving Graph Matching for Partial Face Recognition

Partially Occluded Face Recognition Using Subface Hidden Markov Models

Step-by-step Pipeline Processing Approach for Line Segment Detection.

A Use of Curve Moment Invariants in Recognition of Partially Occluded Objects