Quantisation of Different Color Spaces in Image Retrieval – An Analysis
Aishwarya Harish1, S. Ashwini2, G.S. Anisha3
1Aishwarya Harish, Department of Computer Science and IT, Amrita School of Arts and Sciences, Amrita Vishwa Vidyapeetham, Kochi (Kerala) India.
2S. Ashwini, Department of Computer Science and IT, Amrita School of Arts and Sciences, Amrita Vishwa Vidyapeetham, Kochi (Kerala) India.
3G.S. Anisha, Department of Computer Science and IT, Amrita School of Arts and Sciences, Amrita Vishwa Vidyapeetham, Kochi (Kerala) India.
Manuscript received on 07 April 2019 | Revised Manuscript received on 20 April 2019 | Manuscript published on 30 April 2019 | PP: 1243-1246 | Volume-8 Issue-6, April 2019 | Retrieval Number: F3850048619/19©BEIESP
Open Access | Ethics and 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: We study Content-Based Image Retrieval (CBIR) and in this domain, we compare the performance of different quantized color spaces. This technique is one of the best image retrieval technique that is used worldwide as it produces much better results as compared to the predecessors techniques.CBIR technique makes use of color,texture and shape as the important features for quantization. In our paper, we focus on the color of the images as color has more ability to increase the accuracy of the retrieval. We need to perform queries with images as key. This query image is usually selected from a large image database. The image database that can be used is the Corel’s database (10,000 images).In the first stage of our process, we extract the color features from the image that is the query key; other images present in the dataset are also retrieved; a color descriptor represents the extracted color feature; for this purpose, we use the color histogram. Color histogram helps making the comparison between the images to be more precise. Secondly, the histogram is quantized to reduce computational complexity. The third stage involves the use of distance matrix for similarity measurement. We use Euclidean distance for similarity measure. Currently in wide use are many color spaces such as RGB, HSV, CLE Lab and CLE Luv. We also make a comparative study of how image retrieval performs using RGB and HSV color spaces. In the paper we also provide different tables based on the implementation that clearly helps us to prove which color space has a major play in better retrieval. Based on the implementation tables provided in the below sections, we reach the conclusion that HSV has better image retrieval.
Keyword: Color Histogram, Color Feature Extraction, Euclidean Distance.
Scope of the Article: Image analysis and Processing