1Guangxi Normal University (Yucai Campus,Foreign Language Building,
15 Yucai Road,Qixing District,
Guilin 541004 China - China)
Abstract : Because multi-instance and multi-label learning can effectively deal with the problem of ambiguity when processing images. A multi-instance and multi-label learning method based on Content Based Image Retrieve ( CBIR) is proposed in this paper, and the image processing stage we use in image retrieval process is multi-instance and multi-label. We correspond the instances with category labels by using a package which contains the color and texture features of the image area. According to the user to select an image to generate positive sample packs and anti-packages, using multi-instance learning algorithms to learn, using the image retrieval and relevance feedback, the experimental results show that the algorithm is better than the other three algorithms to retrieve results and its retrieval efficiency is higher. According to the user to select an image to generate positive sample packs and anti-packages, using multi- instance learning algorithms to learn, using the image retrieval and relevance feedback. Compared with several algorithms, the experimental results show that the performance of our algorithm is better and its retrieval efficiency is higher.
https://hal.inria.fr/hal-01383338 Contributor : Hal IfipConnect in order to contact the contributor Submitted on : Tuesday, October 18, 2016 - 2:56:58 PM Last modification on : Thursday, March 5, 2020 - 5:41:05 PM
Chaojun Wang, Zhixin Li, Canlong Zhang. A Multi-instance Multi-label Learning Framework of Image Retrieval. 8th International Conference on Intelligent Information Processing (IIP), Oct 2014, Hangzhou, China. pp.239-248, ⟨10.1007/978-3-662-44980-6_27⟩. ⟨hal-01383338⟩