An effective method to detect seal images from traditional Chinese paintings

At present, more and more traditional Chinese paintings (TCPs) have been digitized and exhibited on the Internet. How to effectively brose and retrieve these images have emerged as a hot topic. Most existing algorithms typically use the global color and texture low-level visual features to describe the content of the traditional Chinese paintings, where the semantic information has been omitted. As an important semantic cue in a traditional Chinese painting, the seal information can provide rich and important semantic meanings to describe the images. To this end, this paper proposed an effective method to detect the seal images from traditional Chinese paintings. The experimental results have demonstrated the effectiveness of the propose method.