MAP-BASED PROBABILISTIC REASONING TO VEHICLE SEGMENTATION

• Published in: Pattern recognition Vol. 31; no. 12; pp. 2017 - 2026
• Main Authors: LEE, JOON-WOONG, KWEON, IN-SO
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.12.1998; Elsevier Science
• Subjects: Algorithmics. Computability. Computer arithmetics; Applied sciences; Artificial intelligence; Computer science; control theory; systems; Exact sciences and technology; Learning and adaptive systems; Pattern recognition. Digital image processing. Computational geometry; Probabilistic reasoning; Segmentation; Theoretical computing; Segmentation; Probabilistic reasoning; MAP; Algorithm performance; Image processing; Reasoning; Maximum likelihood; Expectation; Algorithm; Shape detection; Artificial intelligence; Pattern extraction
• ISSN: 0031-3203; 1873-5142
• DOI: 10.1016/S0031-3203(98)00027-2

This paper proposes a segmentation algorithm by means of a probabilistic reasoning to segment moving vehicles in front of a moving vehicle in a road traffic scene. According to the perceptually known facts of a target, we extract image primitives and update a probabilistic expectation for the target to be in an image. Since a noise image produces unreliable features and degrades the detection and localization, selecting the image primitives, which are less sensitive to noise and represent the facts well, is important. The probabilistic reasoning overcomes this problem baased on MAP ( maximum a posteriori) probability that combines the prior and likelihood probabilities of image features using Bayes’ rule.