Purpose: Partial volume effect is one of the most degrading factors in PET imaging quantitation. The aim of the study was to create three dimensional (3D) representation of the recovery coefficients (RCs) taking into consideration lesion size as well as lesion contrast to improve standardized uptake value (SUV) calculations. Materials and Methods: Several phantom studies with fillable spheres have been conducted at significantly wide range of lesion contrast ratios including 3:1, 5:1, 8:1, 10:1, 12:1, 14:1 and 15:1. The phantom studies were then classified into two groups; one for generating a three dimensional function taking into consideration the sphere size as well lesion to background contrast ratio whereas the other group of phantom datawere used to validate the 3D formulation obtained from the first group. A PET segmentation threshold algorithm was generated based on lesion contrast and lesion size. In addition, another four 3D of the RC of the SUV mean and SUV max were formulated taking into account lesion volume (or diameter) and lesion contrast. Validation of the new algorithms has considered both phantom and clinical studies. Results: Volume threshold optimization revealed significant differences of the threshold value required for the various sphere dimensions at any given contrast ratio. A 3D form has been created that is able to individually segment a PET lesion provided lesion contrast and CT volume. Four functional forms were generated for RCs of theSUV mean and SUV max taking into account lesion volume or diameter while being able to employ lesion contrast in the same formalism. Phantom validation and clinical data suggested the comparable results of the different algorithms with an error of less than or equal to 10%. Conclusion: It has been successful to generate 3D mathematical formulation of the SUV recovery coefficients taking intoconsideration the most influential factors including lesion size and lesion contrast. Validation studies in phantom and clinical data were suggestive of the good performance of the new algorithms generated to correct for partial volume effect. However, further studies are underway to ensure the performance of the proposed algorithms in PET lesion well below the sensitive region of the partial volume effect.