Abstract: An adaptive spatio-temporal compression/decompression algorithm for a video image uses a quadtree hierarchy to represent each frame of the image, the quadtree hierarchy including intermediary nodes and leaves. For each leaf a temporal bit is added indicative of whether a significant intensity change occurred between two frames of the video image. If the temporal bit indicates the significant intensity change occurred between frames, then a new intensity value also is added to that leaf. A decoder using a state machine extracts from the resulting compressed data stream a start address, a block size and the new intensity value for each leaf having the significant intensity change between frames to update a display memory with the new intensity value at each pixel within the area defined by the start address and block size.