Down — Sampling of compressed images in the DCT domain

An efficient down-sampling algorithm of DCT (discrete cosine transform) compressed images is presented in this communication. The algorithm operates directly on the compressed data, thus avoiding the need for decompressing, down-sampling in the spatial domain and re-compressing the images. As a result, the quality of the reconstructed images is higher and the computational complexity is lower than similar algorithms appeared in the open literature. Its structure is highly regular, resulting in efficient software and hardware implementations. The algorithm can be used in various applications, such as image and video browsing, video compositing and transcoding, and HDTV to SDTV conversion.