Selecting features from image data

作者: Hung Khei Huang , Bradley Scott Denney

DOI:

关键词:

摘要: SIFT features are selected from an input image. A procedure is applied to the image obtain candidate keypoints. For each keypoint, there calculation of a first Laplacian value (L u ) for pixels in upper Scale Space and second l lower Space, based on position keypoint. keypoint discarded if L c less than or equal either . In case that not discarded, keypoint's strength s calculated, relative change One more keypoints as corresponding strength.

参考文章(36)
Sébastien Morin, Philippe Vimont, Live-action image capture ,(2009)
Ming-Feng Chiang, Pin-Ching Su, Hwann-Tzong Chen, Image processing method and image processing device thereof for image alignment ,(2013)
Rong Xiao, Zhiwei Li, Rui Cai, Lei Zhang, Learning to rank local interest points ,(2011)
Qiong Liu, Hironori Yano, FUJI XEROX CO., LTD., FUJIFILM Corporation, Novel descriptor for image corresponding point matching ,(2009)
Trevor Darrell, Kristen Grauman, Pyramid match kernel and related techniques ,(2007)
David Marimon SanJuan, Arturo Bonnin Llofriu, Tomasz Adamek, Roger Gimeno Hernandez, Efficient scale-space extraction and description of interest points ,(2011)