Content-Based Image Retrieval System for Real Images

Abstract

With the rapid progress of network technologies and multimedia data, information retrieval techniques gradually become content-based, and not text-based yet. In this paper, we propose a content-based image retrieval system to query similar images in a real image database. First, we employ segmentation and main object detection to separate the main object from an image. Then, we extract MPEG-7 features from the object and select relevant features using the SAHS algorithm. Next, two approaches "one-againstall" and "one-against-one" are proposed to build the classifiers based on SVM. To further reduce indexing complexity, K-means clustering is used to generate MPEG-7 signatures. Thus, we combine the classes predicted by the classifiers and the results based on the MPEG-7 signatures, and find out the similar images to a query image. Finally, the experimental results show that our method is feasible in image searching from the real image database and more effective than the other methods.

Assessment of Object Segmentation Techniques for Object Based Image Retrieval

“…Finally small noisy regions are deleted by using morphological operations. In [7],new segmentation using Quaternion Transform technique for salient region detection is proposed.In [8], proposed the interesting hybrid technique of object segmentation from image using JSEG segmentation followed by active contour models. Here the JSEG segmentation provides segmentation of overall image with grouping the dominant color quantization.…”

Content based Retrieval Management Systems in Web Engineering

“…The problem involved in CBIR of remote sensing images was reduced by applying deep features extraction from the last layers of a trained convolutional neural network from deep learning approaches which gave higher performance than feature extraction using shallow methods [19]. Then, the six features of the image such as RGB Color moments, RGB Color histogram, HSV Color histogram, Cooccurrence Matrices, Local Binary Pattern (LBP) and HSV Color moments were selected by RICE algorithm selection model [20,21] which gave high performance. Its effectiveness was improved by the Scale Invariant Feature Transform (SIFT) descriptor which represents the visual gratified of the images and then utilized the distance ratio as the threshold to control the number of matched feature points.…”