A Novel Image Retrieval Model Based on Semantic Information and Probability Rough Set Analysis Method

Image Retrieval is one of the most active research areas in computer vision and pattern recognition. This paper presents an image retrieval technique based on semantic information and probability rough set analysis method to study uncertain systems. The rough set theory defines an objective type of function and does not require extra parameters or models for analysis, and is therefore very advantageous in image retrieval applications. In this paper, a novel algorithm is described which involves rough neurons for the retrieval of images from medical image databases. Principal component analysis is effectively applied for feature extraction and dimensionality reduction of the feature set. The feature set has been discretized as decision tables using the rough set theory and fed into a 2-layer ANN. 5000 samples have been used for the query application, and an accuracy of nearly 95% has been observed. The computation time is also reduced due to a double reduction in the size of the feature set. The experiment results have been compared with conventional SVM technique and the fuzzy driven image retrieval system and exhibited superiority.