Automated Segmentation in Content Based Image Retrieval System for Marine Life Images Using Shape Feature

摘要

Content based image retrieval (CBIR) has been an active and fast growing research area in both image processing and data mining. CBIR Dynamic nature overcomes/complements the drawbacks of Text Based Image Retrieval (TBIR). In CBIR, extraction of feature and feature matching plays important role in performance of retrieval system. Marine species of Malaysia are rich in biodiversity. We designed and evaluated a CBIR system to characterize the species for future research. Challenges of these images are low resolution, translation, and transformation invariant. In this paper we compared shape feature extraction techniques (Pyramid Histogram of Oriented Gradient, Fourier descriptor, Zernike moment) to obtain effective retrieval results. As a preprocessing we did automatic segmentation of the images and compare it with manual segmentation to check the performance of our CBIR system. Our system automatically detects ROI (region of interest) and extracts required information from the image. Our finding proves that segmentation helps in better retrieval results.