A Varied Efficient Approach on Sketch Based Image Retrieval System
M Indu1, Kavitha K V2
1M Indu, M.Tech Student, Department of Computer Science and Engineering, Sree Chitra Thirunal College of Engineering, Thiruvananthapuram (Kerala), India.
2Kavitha K V, Assistant Professor, Department of Computer Science and Engineering, Sree Chitra Thirunal College of Engineering, Thiruvananthapuram (Kerala), India.
Manuscript received on 10 October 2016 | Revised Manuscript received on 18 October 2016 | Manuscript Published on 30 October 2016 | PP: 51-55 | Volume-6 Issue-1, October 2016 | Retrieval Number: A4741106116/16©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: Especially with the vogue of touch screen devices, retrieval of images that match with a hand-drawn query sketch became a highly desirable feature. Since 1990s, query-by sketch has been an extensive study. Due to the lack of effective and efficient matching solutions they are still very challenging. Compared to face recognition, face photo recognition using face sketch is relatively a younger area. The exceptional triumph of search techniques have encouraged to revisit the problem and focused at solving the problem of sketch based image retrieval. To this end, a novel method is presented here which is as follows: for each image in the database feature extraction is carried out and edge correspondence metric is computed which will be stored. Similarly for the query sketch the same steps are repeated. For each value the query sketch searches for a match score. The database image with highest match score is the retrieved match against the query sketch from the face photo database. An optimized algorithm is also incorporated for images that are corrupted by various types of noises. This method can handle non-facial factors such as such as hair style, hairpins, and glasses. During investigation results show that the proposed method outperforms several state of-the-arts in terms of accuracy and running time.
Keywords: Face Sketch Synthesis, Feature Vector, Edge Correspondence Metric, Sketch Based Image Retrieval.
Scope of the Article: Image Security