Although positron emission tomography (PET) has been commonly used in oncology, for radiation therapy treatment planning or the assessment of response to treatment, in view of the fact that the rate of glucose metabolism is increased in malignant tumours, it is not easily to delineate the boundaries of tumours from the surrounding normal tissue because of the low spatial resolution and inherent noisy characteristics in PET images. In this work, a novel method for automatic segmentation using an active contour model is presented. The algorithm incorporates histogram fuzzy C-means clustering and localized and textural information to constrain the active contour to detect boundaries in an accurat e and robust manner. Moreover, t h latt de Boltzmann method is used as an alternative approach for solving the level set equation to make it faster and suitable for parallel programming. The proposed method was compared with the contourlet-based active contour algorithm and Schaefer's thresholding method using phantom and clinical studies. Our results demonstrate that the developed novel PET segmentation algorithm is applicable to various types of lesions and is capable of producing accurate and consistent target volume delineations, potentially resulting in reduced intra-A nd inter-observer variability commonly observed when using manual delineation.