Abstract: Due to the intensity inhomogeneity and fuzzy boundary of the solid components in partial ground glass opacity ( pGGO) pulmonary nodules，it is difficult to obtain accurate segmentation results through the traditional active contour model． Thus an improved wavelet-energy guided active contour model was proposed for solid component segmentation in pGGO． First，the grayscale information of the image was converted into wavelet coefficients by wavelet transform． The low-pass coefficients were fuzzified to suppress over-enhancement and under-enhancement regions． By integrating high-pass wavelet coefficients and the fuzzified low-pass coefficients，wavelet energy was calculated to build the region term of the active contour model to enhance the dissimilarity between solid part and surrounding ground glass part． Then， the Gaussian mixture model was used to calculate the posterior probability of the pulmonary nodule image． When the posterior probability difference is selected as the boundary detection function， the boundary detection function tended to zero at the boundary of the solid component in pGGO，and the proposed active contour curve stopped evolution． The experimental results show that the model proposed in this paper has a true positive rate of 0. 95，a false positive rate of 0. 23 and a similarity of 0. 80，which contributes to the determination of the solid components in pGGO． 

Key words: pGGO, wavelet energy, active contour model, posterior probability, image segmentation