Robust pattern recognition under shearing and perspective distortions based on Radon transform and non-linear matching

2016 Fiscal Year Final Research Report

• Project/Area Number: 26330206
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Tokyo Denki University
• Principal Investigator: Hasegawa Makoto 東京電機大学, 工学部, 教授 (80303171)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 物体認識 / 文書認識 / せん断歪み / 射影歪み / ラドン変換 / 非線形マッチング / RANSACアルゴリズム

Outline of Final Research Achievements

Robust pattern recognition methods based on Radon transform under various distortions are discussed. Our proposed recognition methods are invariant for Affine distortions on shapes (rotation, scaling, translation, and shearing) using Radon transform and a non-linear matching (Dynamic Time Warping). Moreover, for document recognition, a correction method for perspective distortions using vanishing points made by text lines is discussed. Some noisy lines unrelated to the vanishing points are removed using RANSAC algorithm in the Radon domain.

Free Research Field

情報工学