Techniques for generating 3D images using multi-resolution camera clusters are described. In one example embodiment, the method includes, disposing a multi-resolution camera set including a central camera, having a first resolution, and multiple camera clusters, having one or more resolutions that are different from the first resolution, that are positioned substantially surrounding the central camera. Images are then captured using the camera set. A high resolution depth map is then computed using a hierarchical approach on the captured images. The 3D image of the captured image is then generated using the computed high resolution depth map.