Image Distractor Detection and Processing

In a digital medium environment for distractor removal from an image, a system comprising:

• a distractor processing system, implemented by one or more computing devices, to remove distractors from the image, the distractor processing system including:
• a segmentation module implemented at least partially in hardware to form a plurality of segments from the image;
• a segment prediction module implemented at least partially in hardware to perform operations comprising:
• calculating a score for each of the plurality of segments that is indicative of a relative likelihood that a respective said segment is considered a distractor within the image, the calculation performed using at least one distractor model trained using machine learning;
• classifying the plurality of segments as comprising salient objects that are less likely to be considered important to a user and salient objects that are more likely to be considered important to the user based on the calculated scores for the respective segments;
• outputting the classifications for the plurality of segments;
• and
• a distractor removal module implemented at least partially in hardware to modify at least one segment of the plurality of segments based on the classification of the at least one segment.

In a digital medium environment for training a distractor model to be used for distractor removal in images, a system comprising.

• means for receiving a plurality of images having tagged distractor locations;
• means for calculating values for a set of features for respective ones of a plurality of segments within respective ones of the plurality of images;
• and
• means for generating the distractor model using one or more machine learning techniques in which the distractor model is trained for detection of the distractors using one or more machine learning techniques with the calculated values and the plurality of images.