Depth Analysis of Single View Image Objects based on Object Detection and Focus Measure

In today's Computer Vision era, one important step is 3D scenes generation from 2D images. One of the challenging and difficult tasks is generation of Depth Map from Single View Image. Depth Map gives depth information of pixels in 2D form. Image may have many objects in it and focus measure is one of the important depth cues. In this paper, indoor images with different objects have considered and the focus measures of various objects in the images have calculated using Discrete Wavelet Transform (DWT) by considering it as one of the evaluation metrics. The results show that if the particular object is near, far from the camera, or from one another, the focus parameter changes with respect to that. The object that is near to the camera has greater focus measure as compared to the objects, which are relatively farer. In addition to that, the entropies of different objects computed as another evaluation metrics. Deep Learning used here for object detection from image and the depth map generated.