Automatic word assignment to images based on image division and vector quantization

We propose a method that relates images and words. This method is based on statistical learning from image databases with words. The method uses two processes. The rst uniformly divides each image into sub-images. With this division, all words assigned to images are inherited by each sub-image. The second process clusters sub-images by vector quantization. These processes produce results which show that each sub-image can be correlated to a set of words, each of which is selected from words assigned to original images. After clustering, the voting probability of each word for a set of divided images is estimated. This is done for each cluster of the feature vector of sub-images. Experiments show that this method is e ective.