Performance Enhancement of Content based Image Retrieval System by Integration of Suitable Image Features

Background/Objectives: Retrieving visually similar images from image database needs high speed and accuracy. Various text and content based image retrieval techniques are being investigated by the researchers in order to exactly match the image features. Methods/Statistical Analysis: The proposed technique deals with Content Based Image Retrieval (CBIR) system by applying various image feature detection and matching techniques to investigate the retrieving efficiency of the image. The newness in the proposed research is that it extracts image features like texture, color and shape by combining various angular rotation and image segmentation techniques. With the output of the rotation and segmentation algorithm, color histograms, straight line and outline features including vertical, angled and horizontal lines are extracted. Findings: The accuracy and performance related issues listed in the previous works are overcome while the database images grow. The results obtained by the proposed system obviously confirm that Partitioning and rotating of image objects helps in retrieving many numbers of similar images from the database. Conclusion/Improvements: The proposed CBIR method is compared with our previously existed methodologies and found better in the retrieval accuracy. The retrieval time and accuracy are comparatively good than previous works proposed in CBIR system.