Shape-based multi-spectral optical image reconstruction through genetic algorithm based optimization

Multi-spectral optical imaging of skin and skin-lesions has been of significant interest for various biomedical applications. A multi-spectral optical imaging instrument called Nevoscope has been developed to acquire images of skin and skin-lesion through trans-illumination. Nevoscope based multi-spectral trans-illumination imaging is aimed at characterization of skin-lesions for early diagnosis of skin-cancers by reconstructing distributions of melanin and blood. Conventional approaches usually involve dividing the field of view into a number of voxels and assuming constant optical properties in each voxel. The optical properties are reconstructed in terms of measurements at multiple wavelengths and the distribution of melanin and blood are subsequently calculated. However, since the inverse problem is generally an ill-posed and under-determined one, it is hard to get quantitatively accurate reconstruction. In this paper, a shape-based multi-constrained reconstruction algorithm is presented, which uses a genetic algorithm based optimization methods to find the best possible reconstruction solution. A skin-lesion such as melanoma is modeled as melanin and blood parts, which are delineated by two cubic tensor-product B-spline surfaces. This reduces the number of unknowns to a few control parameters of the surfaces. The parameters are then coded into a genetic algorithm to find a solution through global optimization. Reasonable constraints are incorporated into the genetic algorithm to stabilize the solution. Results of reconstructions of melanin and blood parts are presented for simulated lesions using multi-spectral wavelengths at 580 nm and 800 nm.