ISSN: 2277-9655 [Pinjarkar* et al., 5(12): December, 2016] Impact Factor: 4.116 IC™ Value: 3.00 CODEN: IJESS7 http: // www.ijesrt.com © International Journal of Engineering Sciences & Research Technology [66] IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY TRADEMARK IMAGE RETRIEVAL USING LOW LEVEL FEATURE EXTRACTION IN CBIR Latika Pinjarkar*, Manisha Sharma, Smita Selot * Department of I. T, Shri Shankaracharya Technical campus, Bhila (C.G.),India Department of ETC, Bhilai Institute of Technology, Bhilai(C.G.),India Department of MCA, Shri Shankaracharya Technical campus, Bhilai(C.G.),India DOI: 10.5281/zenodo.188576 ABSTRACT Trademarks work as significant responsibility in industry and commerce. Trademarks are important component of its industrial property, and violation can have severe penalty. Therefore designing an efficient trademark retrieval system and its assessment for uniqueness is thus becoming very important task now a days. Trademark image retrieval system where a new candidate trademark is compared with already registered trademarks to check that there is no possibility of resemblance, has been identified as a major application area of Content Based Image Retrieval (CBIR). This paper proposes an automated system for rotation, translation and scale invariant color trademark recognition. Low level feature extraction techniques used in CBIR are proposed for trademark image retrieval. The performance of the trademark image retrieval system is evaluated using precision and recall. KEYWORDS: CBIR, Trademark, Color feature Extraction, Texture feature Extraction, Shape feature Extraction INTRODUCTION CBIR or Content Based Image Retrieval is the retrieval of images based on visual features such as color, texture and shape. The features of the images in the database are extracted. These features are compared with the features of the query image for similarity computation. The main steps in CBIR are Feature Extraction, Indexing and Retrieval. Feature Extraction – Features are extracted from the images. The description of features is predefined, such as color, texture and shape. These features are saved in the form of real-valued multi-dimensional vectors. Indexing – The database of the extracted image features is then arranged by using an indexing structure for retrieval. Retrieval – The retrieval process comprise of the matching for similarity of the query image with the database images. Content-based retrieval can be done on the indexing structure effectively. Number of applications of CBIR system exists such as trademark or logo image retrieval, Architectural and engineering design, Fashion and interior design and Medical diagnosis etc. Which leads to the big scope of research for improvements in the area of CBIR for the researchers. This paper describes the method for retrieval of Trademark images by implementing feature extraction techniques for Color Feature extraction, Texture feature Extraction and Shape feature Extraction. The rest of the paper is organized as follows: Related work in the area of trademark image retrieval is discussed in section 2. The proposed methodology is described in section 3. Section 4 describes about classification. The experiments and performance evaluation are discussed in section 5. Section 6 concludes the paper.