Combining Focus Measures through Genetic Algorithm for Shape from Focus

For the reconstruction of three-dimensional (3D) shape of microscopic objects different focus measure operators have been employed. It is difficult to compute accurate depth map using a single focus measure due to different type of texture. Moreover, real images with diverse types of illumination and contrast lead to the erroneous depth map estimation through a single focus measure. So to address this problem, we have used spatial focus measure operators in conjunction with genetic algorithm for accurate depth map estimation. Genetic algorithm uses the output of the focus measure operators and using weight updating method accurately estimates the depth maps of real world objects. The performance of the developed method is then evaluated by using both the synthetic and real world image sequences. The experimental results show that the proposed method is more useful in computing accurate depth maps as compared to the existing SFF methods.