Novel Framework of Semantic Based Image Reterival by Convoluted Features with Non-Linear Mapping in Cyberspac
Virender Singh1, Rajeev Gupta2
1Virender Singh, Research Scholar, MMICT & BM, MMDU, Mullana, Ambala (Haryana), India.
2Rajeev Gupta, Assistant Professor, MMICT & BM, MMDU, Mullana, Ambala (Haryana), India.
Manuscript received on 18 June 2019 | Revised Manuscript received on 11 July 2019 | Manuscript Published on 17 July 2019 | PP: 939-942 | Volume-8 Issue-1C2 May 2019 | Retrieval Number: A11610581C219/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In Image Information Retrieval System, finding an effective image is challenging problem. In this problem, main challenge is researcher’s work on texture, color and shape but they ignore semantic relation between these features. In some research emphasis on region base semantic which not only reduces the precision of image retrieval. But region base semantic features uses low-level features and high-level semantic concept to increase the ambiguity in features. In proposed approach find the low-level features and non-linear mapping, to find that use convolution and mapping by poling concept. In polling, parts combine the features of three different layers and find semantic relation by polling concept which reduces the ambiguity of features. In experiment, our emphasis on fruit image retrieval and comparison with color base and region base semantic features approaches but proposed approach shows significant improvement in precision, recall and f-score.
Keywords: Framework Image Semantic Reterival Region.
Scope of the Article: Patterns and Frameworks