This paper presents a Spatial Pyramid image representation-based technique with global and local features captured through DCT coefficient at various levels for offline signature verification. The spatial pyramid feature vector is an extension of an orderless bag of features. We employed a spatial pyramid with 4 levels and took the entire signature image at the first level. The rest of the 3 levels in the pyramid consist of images with 4 partitions, 16 partitions, and 8 \(\,\times \,\) 8 non-overlapping blocks. Our approach captures both local and global DCT features from the image and its sub-blocks at various levels. The AC DCT coefficients obtained are represented in the form of a matrix. The variation in the statistical properties of AC coefficients of the image is made use of in detecting the forgery. For that, the standard deviation and count of non-zero DCT coefficients corresponding to each row in the DCT matrix are found. The standard deviation of DCT coefficients is a good measure of representing the spread of values in an image. The extracted feature vector consists of the standard deviation and the number of non-zero values in the DCT matrix. The Support Vector Machine (SVM) is used for the classification. The classification accuracy of our approach on standard datasets is calculated and the results are compared with a few well-known approaches. This shows that the performance of the proposed approach is better than the other approaches in the state-of-the-art literature.