Title :
Receptive Field Based Image Modeling Method for Interactive Segmentation
Author :
Yang, Bin ; Zhao, Qi-Yang ; Zhang, Rui ; Yin, Bao-Lin
Author_Institution :
Nat. Lab. Software Dev. Environ., Beihang Univ., Beijing, China
Abstract :
In current interactive segmentation algorithms, image models are constructed and simplified to be independent of spatial features of images. This conflicts with receptive field hypothesis of human vision systems, and causes oversegmentation and under-segmentation. Based on receptive field hypothesis, the paper establishes an image modeling method in which spatial distances are taken into account, and a conservative factor is introduced into the image energy function to improve the segmentation veracity. It is shown by experiments that the method is more accurate than its counterparts.
Keywords :
computer vision; image segmentation; interactive systems; human vision systems; image energy function; interactive segmentation; oversegmentation; receptive field based image modeling; receptive field hypothesis; segmentation veracity; spatial features; undersegmentation; Clustering algorithms; Data mining; Data models; Humans; Image edge detection; Image segmentation; Machine vision; Optimization methods; Pixel; Programming;
Conference_Titel :
Image and Signal Processing, 2009. CISP '09. 2nd International Congress on
Conference_Location :
Tianjin
Print_ISBN :
978-1-4244-4129-7
Electronic_ISBN :
978-1-4244-4131-0
DOI :
10.1109/CISP.2009.5303490
