Methods and systems for mask generation for an image or video are disclosed. In one embodiment, the method for creating a mask associated with an image includes receiving a first input of pixels of the image to be included in the mask. The method further includes receiving a second input of pixels of the image to not be included in the mask. The method also includes creating a classifier to classify each pixel of the image as included in the mask, not included in the mask, or unsure. The method also includes training the classifier using the first input of pixels and the second input of pixels. The method further includes classifying each pixel of the image as included in the mask, not included in the mask, or unsure.