In an image, estimating the distance between objects and the camera by using the blur in the images as clue, also known as depth from focus/defocus, is essential in computer vision. However, model-based methods fail when texture-less surfaces are present, and learning-based methods require the same camera settings during training and testing. Now, researchers have come up with an innovative strategy for depth estimation that combines the best of both the worlds to solve these limitations, extending the applicability of depth from focus/defocus.