Abstract: An image compression and restoration technique may be used as a substitute for JPEG. In this technique, orthogonal transforms are performed on the original image signals without interpolation, which means it requires less calculation than JPEG standard. Furthermore, the degradation of the image quality can be suppressed. After that quantization, re-ordering and Huffman coding follow. The Huffman coding is also unique in the sense that unlike JPEG, it uses the data from the previous blocks which results in a more simple, less CPU intensive codec technique. Conversely, the compressed data is restored using a procedure reverse to that of the aforementioned compression technique.

Abstract: An improved motion vector detecting method is provided that can ensure a wide search area while suppressing detection of erroneous motion vectors at the same time. Specifically, when detecting motion vectors from a first and a second image frame, the method divides the second image frame into blocks, classifies the blocks into a first and a second group such that the mutually adjacent blocks belong to the different groups. The method specifies a motion vector candidate for each of the blocks in the first and second groups, then performs the process of updating the motion vector candidate by studying an average of the motion vector candidates of the adjacent blocks for each of the blocks in the first and second groups. The process is repeated for a certain time.