Pictogram Extraction Algorithm for Local and Parallel Processing

Abstract

Keywords: image processing, pictogram extraction, scenery image, local and parallel image processing, support for sight-restricted peopleIn this paper, we propose a pictogram extraction algorithm for local and parallel processing. Then, we show the example that the proposed algrithm is applied to lavatory mark detection. This algorithm works under a severe constraint: each pixel of a result image must be derived from only information of their neighbor pixels. This constraint is very important for a low cost device like a mobile camera because many other algorithms using grobal process can not be really carried out in such a device. Fig. 1 shows our target device. It consists of many Processing Elements (PE). This device can process images in parallel due to independency of PEs.The proposed algorithm can estimate degree of square by pixel count method to extract pictograms. Fig. 2 shows the flow of the proposed algorithm. Fig. 3 shows examples of computer simulations. The result shows that the proposed algorithm can extract pictogram regions in a rate of more than 90%, and false detection rate is less than 20%.