In this study, we present a fast and robust practical tool for segmentation of solid tumors with minimal user interaction. The lung image is segmented manually by the experts may includes risk and time consuming process. Hence, in this study, the lung image is segmented in order to identify the tumor sector. Initially, the input lung image is applied with the denoising process for removing noises with the aid of multi-wavelet transformation. After this process, the CA algorithm is applied for obtaining the background seeds and foreground seeds (tumor seeds) and then the level set algorithm is applied for acquiring the acute tumor tissues. As a result of the mentioned process, the tumor sector is segmented and the results are depicted. Studies on lung tumor datasets demonstrate 80-85% overlap performance of the proposed algorithm with less sensitivity to seed initialization, robustness with respect to heterogeneous tumor types and its efficiency in terms of computation time.