A Lossless Compression MethodofMedical Imagesbased ontheLocal Texture Feature
Yin Zhang · 2005
As fora kindofmedical imageswithclear texturefeature, a new lossless compressive methodis presented inthispaper. Withtwomainsteps, themethod firstly sets upthepredictive modelofthecurrent pixel withits surrounding texture features, including itsfourcolor change gradients; andthenmeasures thesimilarity ofallpixels within theimagewiththeir predictive modelssothattheredundancy broughtaboutbythesimilarity isremovedanda higher compression ratio isobtained. Thefollowing simulation results showthatthenew methodismuch moreefficient than JPEG-LSalgorithm inthecompression ofsuchkindsof medical images. I.INTRODUCTION Inestimating theeffectiveness ofalossless compression way,thebit/pixel ratioaftercompression isalways considered tobeveryimportant. Forthereason that nodata oftheimages areallowed toloss during thecompression, its efficiency isdifficult toimprove. However, insomespecial application areas, suchasmedical images process, there is anincreasing conflict between theincreasing numberofthe digital images andlimited storage spaceandtransportation bandwidth, sohighefficient lossless compression methods