A unified model-based approach to ATR that uses 3D models to control detection, segmentation, and classification is described. Objects are modeled by rectangular boxes whose dimensions are Gaussian random variables. A fast predictor estimates the size and shape of expected objects in the image, which controls detection and segmentation algorithms. Segmentation fits oriented rectangles (length x width @ pose) to object-like regions detected using a multi‐level thresholding/region tracking approach. Detections are classified by comparing measured to predicted region length and width in the pose direction. The method is fast and requires only a coarse characterization of objects/object classes.

Date Published: 21 May 2015
PDF: 10 pages
Proc. SPIE 9474, Signal Processing, Sensor/Information Fusion, and Target Recognition XXIV, 94740Z (21 May 2015); doi: 10.1117/12.2179810

Published in SPIE Proceedings Vol. 9474:
Signal Processing, Sensor/Information Fusion, and Target Recognition XXIV
Ivan Kadar, Editor(s)