Obstruction removal is a new task in Computer Vision, nowadays, machine can help people to remove somethings make they picture imperfect such as fences, reflection and raindrops. With the convenience of it, we can apply for the reporter to take the picture through the glasses or remove noise in the photos from photographer.
The goal of obstruction removal is to remove 3 mains noise: rain drop, reflection and fence. By alternating between estimating dense optical flow fields of the two layers and reconstructing each layer from the flow-warped images via a deep convolutional neural network, after that learning-based layer reconstruction allow to accommodate potential errors in the flow estimation and brittle assumptions such as brightness consistency, we can remove these three noises.
Obstruction removal with the purpose to remove unwanted object in the images.By leveraging the motion differences between the background and the obstructing elements to recover both layers in order to remove unwanted object and generate full images without obstruction. The positive of this method is helping users to remove unwanted object in the images by using Deep Learning.