A device for removing a noise artifact from a document receives a scan of the document, where the document contains a noise artifact at least partially obstructing a portion of the document. The device generates an image of the document, and extracts a first set of features from the image. The device identifies noise artifact features from the first set of features, and generates a second set of features by removing the noise artifact features. The device generates a test clean image of the document based on the second set of features. The device determines whether a portion of the test clean image that previously displayed the noise artifact corresponds to a counterpart portion of the training clean image. If it is determined that the portion of the test clean image corresponds to the counterpart portion of the training clean image, the device outputs the test clean image.