Content Based Image Retrieval (CBIR) is the process of querying by image content, which can be represented by using color, shape and texture features. These features can extract in both spatial and frequency domain. Many existing systems are using hybrid features by mixing two or more feature types. This papers addresses the feature extraction by combining spatial and frequency domain features. Edge descriptors make the object more clear in the image. Canny edge algorithm is used to sharppen the image in spatial domain. Two kinds of texture features are extracted using uniform Local Binary Patterns (LBP) and wavelet based Haralick texture features. Daubechies 20 is applied on the sharppened image to get the approximation image in the wavelet sub band. The experiment has done on two medical datasets which are publicly available. By using this methodology we can able to build a sophisticated query system for large scale image databases. © 2018, Institute of Advanced Scientific Research, Inc. All rights reserved.