Incremental model JPEG image segmentation decoding method for embedded system

Abstract

The invention discloses an incremental model JPEG image segmentation decoding method for an embedded system. The method comprises the following steps: cutting JPEG image into a plurality of regions, obtaining the positioning information and decoding state information of a region to be decoded, decoding one region segment data in each scan contained in JPEG image, adding together to obtain quantified DCT conversion coefficient corresponding to the region, and obtaining a section of JPEG image corresponding to the region used for display and storage through inverse quantization and IDCT conversion. The incremental model JPEG image segmentation decoding method of the invention can realize the segmentation decoding of Progressive JPEG image so as to greatly reduce the demand for storage space; in addition, data of each scan are not needed to perform inverse quantization and IDCT conversion repeatedly so as to greatly reduce the overall computational quantity.

Description

technical field [0001] The invention relates to an image decoding method, in particular to a progressive mode JPEG image segmentation decoding method used in an embedded system. Background technique [0002] JPEG (Joint Photographic Experts Group) is a national standard (ITU-T T.81) for grayscale or color static image compression. Due to its high compression ratio and good image quality, it has been widely used in the Internet, medical imaging, Various fields such as satellite photography and commercial image system. [0003] For a typical progressive mode (Progressive) JPEG decoding system, it is necessary to allocate a memory buffer consistent with the size of the entire image to save all quantized DCT (Discrete Cosine Transform, discrete cosine transform) after Huffman decoding ) transformation coefficient. The decoder first performs Huffman decoding on the first scan (scan) data, and the obtained quantized DCT coefficients are stored in the memory buffer, and then inve...

AI Technical Summary

Problems solved by technology

As the image size increases, the demand for memory space also increases. For example, a picture with a size of 2048×1440 (about 3 million pixels) requires approximately 8-16MB of memory depending on the encoded YUV format. buffers, which is difficult in memory-constrained applications and completely unacceptable in embedded applications

[0006] (2) For each scan data, the process of Huffman decoding, inverse quantization, IDC transformation, and image superposition must be repeated, and the process of inverse quantization and IDCT (Inverse DCT, inverse discrete cosine transform) transformation takes up a lot of time Computation

Images