310 likes | 660 Views
Face Recognition. Joshua I. Cohen. Overview. Cutting Edge Constraints: - Speed - Accuracy. Data Acquisition. Digital Camera. Data Acquisition Cont’d. Data Acquisition Cont’d. Video Camera w/ IR Filtering. Algorithm. Pre-Processing Convert .bmp to .pgm Isolate Largest Contour
E N D
Face Recognition Joshua I. Cohen
Overview Cutting Edge Constraints: - Speed - Accuracy
Data Acquisition • Digital Camera
Data Acquisition Cont’d • Video Camera w/ IR Filtering
Algorithm Pre-Processing Convert .bmp to .pgm Isolate Largest Contour Profile Curve Curve Matching Parse Data to Extract Costs
I. Pre-Processing Starting Images
I. Pre-Processing Cont’d Convert To Gray Scale
I. Pre-Processing Cont’d Binary Threshold
I. Pre-Processing Cont’d Centroid
I. Pre-Processing Cont’d Filter Noise Using Connected Regions
I. Pre-Processing Cont’d Fill Region Up Until Contour
I. Pre-Processing Cont’d Remove Hair-Line Noise
I. Pre-Processing Cont’d Crop Image Area
I. Pre-Processing Cont’d Normalize Fill Border To Black Save to .bmp File
II. Convert .bmp to .pgm • SUN utility • convert-to-pgm f1.bmp f2.pgm .bmp .pgm
III. Isolate Largest Contour • SGI utility • computeBoundaryLength f1.pgm –smooth 1 CONTOUROPEN/CLOSEnx1 y1 x2 y2 …xn yn .con file
V. Curve Matching • OpenCurveMatch f1.con f2.con 10 1 • Intrinsic Properties: length and curvature • Optimal Correspondence • Alignment Curve • Invariant: rotations, translations, viewpoint variation, articulation, occlusions
VI. Parse Data to Extract Costs • f1-points-f2-points-match.txt • cost-summary.txt f1 f1 0 f1 f2 403.668 f1 f3 468.081 f1 f4 454.315 f2 f2 0
Preliminary Results Amir vs Everyone
Preliminary Results Cont’d Brian vs Ming / Vinesh
Preliminary Results Cont’d Tom vs Vinesh
Preliminary Results Summary 50-60% success rate margin of error very small - ~ 450 or less is a match - ~ greater than 450 not a match
Remaining Issues Angle Eye
Remaining Issues Cont’d • Angle Cont’d Forehead / Eye Normal
Remaining Issues Cont’d • Hair-Line • Normalizing Distortion
Possible Solutions • Use More Criteria (local extrema) • - contour of nose • - contour of chin • Better Method to Filter Hair-Line • Curve Match Segments of Face Profile
References • Ross Cutler, “Face Recognition Using Infrared Images and Eigenfaces”, April 1996. • Thomas B. Sebastian, Philip N. Klein, Benjamin B. Kimia, “Alignment-based Recognition of Shape Outlines”, 2001. • Zdravko Liposcak and Sven Loncaric, “A Scale-Space Approach to Face Recognition from Profiles,” 1999.