作者: Yoong-Chert Foo , Yu-Ling Chen , Kai Huang , Peter Leslie Bell , Michael John Webb
DOI:
关键词:
摘要: Disclosed is method of detecting a predetermined mark (30) forming part an image. A digitized form (48, 49) the image (21) comprising plurality picture elements provided and within current region set features determined. The typically comprise dominant color feature (44) region; at least one moment (111) based on pertaining to region. Each determined compared (113) with corresponding (114) mark. Using comparison, confidence level for detection said calculated. Based level, whether or not detected in can then be Adjunct above method, also disclosed are methods providing multiple encoded representations (40) image, center location (108, 160) pattern determining (170, 171) portion