Compression of aerial images for reduced-color devices Pasi Fränti and Ville Hautamäki Department of Computer Science, University of Joensuu Box 111, FIN-80101 Joensuu, FINLAND ABSTRACT Mobile devices are more often capable for locating the user on the globe by using GPS or network-based systems. The location is given to the user in a meaningful context such as map or aerial image. We study compression methods for reducing the amount of data required by aerial images. We consider the two well known lossless graphics file formats GIF and PNG, the compression standards JPEG and JPEG2000, a fractal compressor known as FIASCO, and commercial wavelet-based method developed for aerial images. It is also expected that the devices support fewer than 256 gray levels. It is therefore possible to get further reduction by quantizing the images prior to compression, for example, to two-bit per pixel, and then apply lossless compression methods such as JBIG, GIF and PNG. For four color mobile displays error diffusion dithered images approximate the original 8-bit color images quite well. The trade-off in dithering is that the lossless compression ratios decrease. Solution to this might be to store the 8-bit images compressed by using a decent lossy compressor such as JPEG2000. Quantization and dithering should happen only at the moment when the image is displayed. Keywords: Aerial images, image compression, mobile imaging, JBIG, JPEG, JPEG2000. 1. INTRODUCTION Aerial images are photographs of terrain taken from air or from satellite. Conventional maps can be represented in digital form as vector graphics or as bi-level bitmap. Aerial images, on the other hand, need more color information. The mobile devices, however, have limited storage and reproducing capabilities. It is therefore important that the image can be compressed efficiently. In this report, we study different compression methods for reducing the amount of data required by aerial images. The original images are 8-bit per pixel (bpp) gray-scale images, which give plenty of choices for the compression methods. We consider the two well known lossless graphics file formats GIF [4 ] and PNG [20 ], the current compression standards JPEG [21 ] and JPEG2000 [10 ], and a fractal compressor known as FIASCO [19 ], and commercial wavelet-based method ECW developed for aerial images [16 ]. Impact of the source data suitability for different compression methods can be seen especially from poor performance of lossless GIF. The PNG method is supposed to achieve always slightly better results than the popular GIF. In this case, GIF produced unacceptable larger-than-original compression ratio, while PNG files were about 2/3 of the size of GIF. In lossy category, wavelet-based JPEG2000 clearly outperformed JPEG, FIASCO and ECW by a wide margin. For four color mobile displays error diffusion dithered images approximate the original 8-bit color images quite well. The trade-off in dithering is that the lossless compression ratios decrease. Solution to this might be to store the 8-bit images compressed by using a decent lossy compressor such as JPEG2000. Quantization to 2-bit and dithering should happen only at the moment when the image is displayed. This approach is demonstrated in Figure 1. The choice of the compression method always heavily depends of the details of source data, as is also seen in this case. If the source data would have been pre-compressed with different method, or not compressed at all, or if it had different noise factors than the present artifacts from JPEG method, compression results might be slightly different. Nevertheless, the JPEG2000 was still the best method for 8-bit images and possibly also for 2-bit images. The rest of the report is organized as follows. In Section 2, we briefly recall the compression methods and give pointers where sources can be found. In section 3, the test image set is described. Compression results with 8-bit and with 2-bit images are summarized in Section 4. Conclusions are drawn in Section 5.