A robust solution for object recognition by mean field annealing techniques

✍ Scribed by Jung H. Kim; Sung H. Yoon; Chil-Woo Lee; Kwang H. Sohn

• Book ID: 104161357
• Publisher: Elsevier Science
• Year: 2001
• Tongue: English
• Weight: 405 KB
• Volume: 34
• Category: Article
• ISSN: 0031-3203
• DOI: 10.1016/s0031-3203(00)00024-8

✦ Synopsis

Object recognition in multi-context scene is one of the very di$cult problems to "nd a robust solution in many applications. The annealed Hop"eld networks have been developed to "nd global solutions of a non-linear system. In the study, it has been proven that the system temperature of MFA is equivalent to the gain of sigmoid function of Hop"eld network. In our early work, we developed the hybrid Hop"eld network (HHN) on the purpose of fast and reliable matching in the object recognition process. However, HHN does not guarantee global solutions and yields false matching under heavily occluded conditions because HHN is depending on initial states by its nature. In this paper, we present the annealed Hop"eld network (AHN) to "nd a robust solution for occluded object matching problems in multi-context scenery. In AHN, the mean "eld theory is applied to the hybrid Hop"eld network in order to improve computational complexity of the annealed Hop"eld network and provide reliable matching under heavily occluded conditions. AHN is slower than HHN. However, AHN provides near global solutions without initial restrictions and provides less false matching than HHN. The robustness of the algorithm is proved by identifying occluded target objects with large tolerance of their features. Also, we present a optimal boundary smoothing algorithm to extract reliable features from the boundary representation of the object heavily contaminated by noise.