Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/311441
Title: | Design and Implementation of Improved Content Based Image Retrieval System In Trademark Registration |
Researcher: | Pinjarkar, Latika |
Guide(s): | Sharma, Manisha and Selot, Smita |
Keywords: | Computer Science Computer Science Information Systems Engineering and Technology |
University: | Chhattisgarh Swami Vivekanand Technical University |
Completed Date: | 2019 |
Abstract: | Content-based image retrieval (CBIR) is the foundation of image retrieval systems now days. For obtaining more accurate retrieval results, Relevance Feedback (RF) approaches were integrated with CBIR by taking into account the user s feedbacks information. newlineTrademark recognition and retrieval is a vital application component of Content Based Image Retrieval (CBIR). It deals with matching of the input trademark or logo with stored trademark images in database. This application, under CBIR umbrella, focuses on optimizing search through database by extracting minimum features from set of the images and using relevance feedback mechanism to identify the relevant images. Researchers working in the field of trademark image retrieval have implemented the approaches like quantized representation of the logo/trademark regions, bundling the local features and the features from the spatial neighborhood of trademark images in one unit and learning a statistical model for the distribution of wrong detections. Recently the deep learning approach is also employed by the researchers for trademark image retrieval and the retrieval results achieved in terms of mean average precision were 74.4% ( Iandola et al.(2015)) and 84.2% (Bao et al. (2016)). While Iandola et al. (2015) achieved an accuracy of 89.6%. newlineReduction in semantic gap, reduction in computation complexity and hence in execution time and attaining more accuracy are the major challenges in designing and development of trademark retrieval system. newlineTill date nobody has addressed these issues by integrating optimization and/or clustering approaches with the Relevance Feedback mechanism. Also the integration of deep CNN with Relevance Feedback has not been observed for trademark image retrieval. None of the researchers have suggested new similarity metric for computing the similarity between query image and the database images of trademark. Also new clustering technique in the field of trademark image retrieval has not been proposed till date. newlineThe direction of the proposed wor |
Pagination: | 7p.,140p. |
URI: | http://hdl.handle.net/10603/311441 |
Appears in Departments: | Department of Computer Science and Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 81.6 kB | Adobe PDF | View/Open |
02_certificate.pdf | 415.04 kB | Adobe PDF | View/Open | |
03_preliminary_pages.pdf | 253.8 kB | Adobe PDF | View/Open | |
04_chapter1.pdf | 287.38 kB | Adobe PDF | View/Open | |
05_chapter2.pdf | 301.16 kB | Adobe PDF | View/Open | |
06_chapter3.pdf | 587.61 kB | Adobe PDF | View/Open | |
07_chapter4.pdf | 374.61 kB | Adobe PDF | View/Open | |
08_chapter5.pdf | 1.3 MB | Adobe PDF | View/Open | |
09_chapter6.pdf | 180.55 kB | Adobe PDF | View/Open | |
10_references.pdf | 293.85 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 255.16 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: