Development of High-Accuracy Image Matching Techniques Using Feature Descriptors Based on Local Phase Array

2017 Fiscal Year Final Research Report

• Project/Area Number: 15H02721
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Tohoku University
• Principal Investigator: Aoki Takafumi 東北大学, 情報科学研究科, 教授 (80241529)
• Co-Investigator(Kenkyū-buntansha): 本間 尚文 東北大学, 電気通信研究所, 教授 (00343062) ; 伊藤 康一 東北大学, 情報科学研究科, 助教 (70400299)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 画像，文章，音声等認識 / ディジタル信号処理 / 画像照合 / 位相限定相関法 / 計算機システム

Outline of Final Research Achievements

This project developed a new feature descriptor called local phase array, which is defined by phase information of Fourier transform of given images, and its advanced techniques. We applied developed techniques to a wide range of applications such as robust matching of biometric images, multi-view stereo image measurement, fast and high-accuracy machine vision, image measurement for airborne SAR images and medical volume data analysis. We also implemented developed methods on GPUs and FPGAs. We demonstrated the effectiveness of local phase array through its practical applications.

Free Research Field

計算機科学