Legible Picwords: Render a Picture by Packing Keywords

Abstract

When people see the picture of a face , they might want to get more information about it. Since the image itself cannot tell the specifics, other sources of information, for example scripts from the internet, may need to be used. A system can be developed to create an image that contains face and information. PicWords is a regular arrangement of both picture and keywords. It is an existing system and mosaicking and calligram are the related works for it. Mosaicking is one of the techniques of image processing for joining together several arbitrarily formed images to form one large balanced image so that the borders between the original images are not seen. Calligram is a modification of a large target image by placing a collection of small words; each precisely selected to fit a specific target image block. Viewer is not able to read the content related to the target image in both cases. In a PicWord system, a two-part binary image, called as silhouette is produced from the source image. The first part can be used for placing the keywords, while the second part can be ignored. The first part is then segmented into small patches, each of which serves as a single keyword container. To ensure that more significant keywords are placed in larger and more prominent image patches, both the patches and keywords are classified and created a correspondence between the patch list and the keyword. Then keyword-patch warping is done for deforming the text. Symbol filling technique is used to improve the beauty of PicWords. In the existing PicWord, the words are not clearly understand from the image, so the legibility is less. The objective of this paper, is to develop an image associated with texts content system called PicWords with improved legibility.