Binary Trademark Image Retrieval Using Region Orientation Information Entropy

This paper presents a new trademark image retrieval method based on the region orientation information entropy. In the first stage, image is rotated according to principal orientation, and the object region in the rotated image is extracted. Then, the object region is partitioned into a lot of sub-blocks. In the third stage, the information entropy of each partitioned region is computed, which construct a feature vector for describing the shape of the image. Finally, the Euclidean distance is adopted to measure the similarity between the images based on the feature vector of each image obtained. Experiments performed on a database containing more than 2,000 trademark images show that the region orientation information entropy keeps good invariance under rotation, translation, scale and the retrieved results satisfy human visual perception very well.