Image Content Based Retrieval System using Cosine Similarity for Skin Disease Images Sukhdeep Kaur 1 , Deepak Aggarwal 2 1 Computer Science and Engineering, PTU Jalandhar, RBGI, Mohali Campus, Kharar, Punjab 140301, India Sukh.kaur.rb@gmail.com 2 Computer Science and Engineering, PTU Jalandhar,BBSBEC, Fatehgarh Sahib, Punjab 140301, India Deepak.aggarwal@bbsbec.ac.in Abstract A content based image retrieval system (CBIR) is proposed to assist the dermatologist for diagnosis of skin diseases. First, after collecting the various skin disease images and their text information (disease name, symptoms and cure etc), a test database (for query image) and a train database of 460 images approximately (for image matching) are prepared. Second, features are extracted by calculating the descriptive statistics. Third, similarity matching using cosine similarity and Euclidian distance based on the extracted features is discussed. Fourth, for better results first four images are selected during indexing and their related text information is shown in the text file. Last, the results shown are compared according to doctor’s description and according to image content in terms of precision and recall and also in terms of a self developed scoring system. Keyword: Cosine similarity, Euclidian distance, Precision, Recall, Query image. 1. Basic introduction to cbir CBIR differs from classical information retrieval in that image databases are essentially unstructured, since digitized images consist purely of arrays of pixel intensities, with no inherent meaning. One of the key issues with any kind of image processing is the need to extract useful information from the raw data (such as recognizing the presence of particular shapes or textures) before any kind of reasoning about the image’s contents is possible. An example may make this clear. Many police forces now use automatic face recognition systems. Such systems may be used in one of two ways. Firstly, the image in front of the camera may be compared with a single individual’s database record to verify his or her identity. In this case, only two images are matched, a process few observers would call CBIR[15]. Secondly, the entire database may be searched to find the most closely matching images. This is a genuine example of CBIR. 2. Structure of CBIR model Basic modules and their brief discussion of a CBIR modal is described in the following Figure 1.Content based image retrieval system consists of following modules: Feature Extraction: In this module the features of interest are calculated for image database. Fig.1 Modules of CBIR modal  Feature extraction of query image: This module calculates the feature of the query image. Query image can be a part of image database or it may not be a part of image database. Similarity measure: This module compares the feature database of the existing images with the query image on basis of the similarity measure of the interest[2]. Image Database Feature database Feature Extraction Results images Query image Indexing Similarity measure Feature extraction of query image ACSIJ Advances in Computer Science: an International Journal, Vol. 2, Issue 4, No.5 , September 2013 ISSN : 2322-5157 www.ACSIJ.org 89 Copyright (c) 2013 Advances in Computer Science: an International Journal. All Rights Reserved.