Age invariant face recognition using 3 D models. In which you can make a code that will recognize person at age x and age x+15 also

It can be used for 3D pattern recognition.

Processing of medical images to construct volume from CT/MRI slice and volumetric analysis will be interesting

Another project can be extracting the spatial-layout of indoor images with RGB-Depth data.

Object Segmentation and grasp planning in cluttered scenes

Detection of various volumetric features such as edges, wireframe contours, skeletal forms in 3-D digital images

Extracting tumor from 2D brain MRIs and converting them to 3D image