Image Mosaics with Irregular Tiling

We present an approach for generating image mosaics with irregular tiles made up from patches taken from photographs, paintings and texture images. We propose a method to generate irregular tiling patterns using polygon tessellation in conjunction with a feature-based segmentation scheme, so that features in the input image can be better preserved in the generated mosaics. In order to avoid the mismatch in roughness between the sub-image in the tile region of the input image and tile images in the tile dataset that may arise in the previous RGB color based image descriptors, we introduce the concept of region entropy into the image descriptor to achieve better match in both color and roughness. A few mosaic images generated by our system are presented, some of which have effects of Chinese characteristics.