In recent years, researchers have increasinglytargeted the development of hybrid systems to improve thequality of the compressed image. Hybrid methods arecombinations of several techniques that take into account theadvantages and disadvantages of it, in order to obtain a betterimage quality with a high compression ratio in a minimalcalculation time. In this article, we are interested in geometricactive contours techniques using the level-set method and thequincunx wavelet , so we proposed an hybrid algorithm for colormedical image compression based on the level set method andthe quincunx wavelet transform coupled with the setpartitioning in hierarchical trees (SPIHT) algorithm.Experimental results show that the proposed algorithm givesbetter results compared to various methods, where it providesvery important PSNR and MSSIM values