In this paper, a novel hashing framework is proposed using edge map and image normalization. In the proposed method, the input image is normalized using geometric moments and an edge map is then employed using the canny edge detector. The estimated binary image is divided into non-overlapping blocks and a chaotic map is used for random block selection. The singular value decomposition is carried out on the selected blocks for extracting the significant features followed by generation of hash value. The simulation results support the contention that the proposed technique is secure and considerately robust against a variety of image manipulations. © 2021, Springer Nature Singapore Pte Ltd.