IMAGE PROCESSING DEVICE, IMAGE RECOGNITION DEVICE, IMAGE PROCESSING PROGRAM, AND IMAGE RECOGNITION PROGRAM
作者： Enokida Shuichi , Yamada Hideo , Kuno Kazuhiro , Shibata Masatoshi , Ohtsuka Hiromichi
DOI:
关键词:
摘要: The present invention reduces calculation costs while raising detection accuracy. An image processing device 8 has a function for plotting luminance gradient co-occurrence pair of an on feature plane 15 and applying EM algorithm thereto to form GMM. learns pedestrian creates GMM the image, subsequently background furthermore calculates difference between two generates relearning based calculation. plots sample that conforms by inverse theorem. finally forms represents distribution samples at designated mixed number thereby standard serves as recognition. When this is set less than earlier, dimensions with which analyzed are reduced, making it possible reduce costs.
lens.org 参考文章(6)
Sohei Iwata, Shuichi Enokida, Object Detection Based on Multiresolution CoHOG international symposium on visual computing. pp. 427- 437 ,(2014) , 10.1007/978-3-319-14364-4_41
Daisuke Aoki, Junzo Watada, Human tracking method based on improved HOG+Real AdaBoost asian control conference. pp. 1- 6 ,(2015) , 10.1109/ASCC.2015.7244780
Yuji Yamauchi, Chika Matsushima, Takayoshi Yamashita, Hironobu Fujiyoshi, Relational HOG feature with wild-card for object detection international conference on computer vision. pp. 1785- 1792 ,(2011) , 10.1109/ICCVW.2011.6130465
Haoyu Ren, Ze-Nian Li, Object Detection Using Generalization and Efficiency Balanced Co-Occurrence Features international conference on computer vision. pp. 46- 54 ,(2015) , 10.1109/ICCV.2015.14
Hironobu Fujiyoshi, 弘亘 藤吉, Masanari Takagi, 雅成 高木, Image recognition apparatus ,(2007)
Hironobu Fujiyoshi, 弘亘 藤吉, Object detecting apparatus ,(2008)