Performance Analysis on Discrete Wavelet and Curvelet Transform in Histology Image Retrieval

Abstract

Content Based Image Retrieval (CBIR) plays a major role in decision making related to clinical activities. CBIR systems differ by the way of extracting features. Many feature extraction methods such as Zernike moments, Tamura, Gabor Texture, Scale-invariant feature transform, discrete cosine transform and Gray-level co-occurrence matrix are in practice. This paper proposes two feature extraction methods that use Curvelet and Wavelet Transforms. Further coefficient of variations is used in matching the similar images. Experimental results show that the proposed method is efficient in terms of precision and recall. Results shows that the proposed method is well suitable for histology images.