Dpcm With Integer Wavelet Transform Infrared Medical Image Compression Method

Infrared thermal image technology, based on infrared radiation theory, is image diagnosis technology to examine temperature distribution of human body. Due to its no injury, no damage, no radiation, non-contact and convience, it has being widely used in clinical medicine and scientific research. With the development of modern computer technology and various medical imaging technologies, digital medical picture data increases greatly, leading to more lack of limited storage space and the restraint of transmission bandwidth. Infrared medical picture is special in that wrong description costs a lot and a compression of high reconstruction quality has to be provided. Without effective compression, the correct medical diagnosis can't be guaranteed and will impede the development of Hospital Information system (HIS), Picture Archives and Communication System (PACS) and long-distance medical treatment.Based on the features of infrared medical pictures, this paper adopts the following methods to realize the approximate lossless compression of the infrared medical pictures.1, Removal of background: Based on the fact that the background temperature is lower than the normal body temperature, we first use edge diction to get the edge of the human body, then we set all the value below the edge temperature to the normal measure temperature so as to the following processing.2, Integerating the source data: The source data are stored in float number which afeects the compression result; therefore, before we compress the data, we must turn the float number into integer number.3 , Adaptive DPCM: In order to reduce the residual among the pixels, we adopt the adaptive DPCM;4, Lifting wavelet transform: In order to realize the lossless compression, we adopt the lifting wavelet transform;5, set partition in hierarchical tree (SPIHT) : we use the famous algorithm :SPIHT to compress the wavelet coefficients;6, Arithmetic coding: Arithmetic entropy coding is used to compress the binary bit stream.Computer analogue experiments show that the compression ratio obtained improves greatly with combination of DPCM and integral small wave transformation, compared with only using a single method of them; and good signal-to-noise ratio and reconstructed picture will be obtained.In brief, the compression method put forward in this paper is a good algorithm, for it is able to improve compressin ratio with approximate no damage to pictures.