A region-labeling approach of image reconstruction under continuous-wave mode is developed for “coarse-grain” diffuse optical tomography (DOT). The method is based on the framework of the pixel-based DOT methodology and on an assumption that different anatomical regions have their respective sets of homogeneous optical parameter distributions. The proposed scheme is validated using a three-dimensional (3D) digital mouse atlas, the optical parameters of the organs could be simultaneously reconstructed at 60 dB, 40 dB and 20 dB noise levels. Physical experiments on a phantom are also conducted to evaluate the performance of the method. The reconstructed results show that the region-labeling DOT solution greatly improves the ill-posedness of the inverse problem as well as the imaging resolution and quantitative accuracy by reducing the number of unknowns to be reconstructed. The homogeneous optical parameters of each region can be reconstructed from noisy data.