Image/video logo recognition based on distance transform
视频帧/图像中图标识别算法

This paper proposed a new recognition and retrieval algorithm based on distance transform for logo in image and video.In order to reduce the influence brought by color,the logo region firstly transformed into binary image.Then introduced an efficient logo image recognition method based on location information of pixel.Unitied the gray-level logo image,and followed binarization.Then rotated it on a main direction,applied distance transform to translate the location information into gray level information.By comparing the histogram feature,logos were coarsely recognized.The experiments halyd done in a database including 1 000 binary logo images.Experiment results show that the algorithm possesses high processing speed and recognition accuracy,and it has very excellent performance under translation,scale and rotation.