Technologies for Automatically Determining and Displaying Salient Portions of Images

A system for automatically determining and displaying salient portions of images, comprising:

• a user interface;
• a memory storing a set of computer-readable instructions comprising at least a saliency detection learning model that is trained with pixel data of a plurality of training images depicting respective features of interest and is configured to output one or more saliency maps corresponding to the respective features of interest;
• and
• a processor interfacing with the user interface and the memory, and configured to execute the set of computer-readable instructions to cause the processor to:
• analyze, by applying the saliency detection learning model, a digital image to generate at least one saliency map corresponding to at least one feature of interest in the digital image, the at least one saliency map selected from the one or more saliency maps of the saliency detection learning model,
• determine one or more salient portions of the digital image at least by:
• generating a threshold image based on one or more edges of the at least one saliency map,
• determining, based on the threshold image, one or more minimal bounding fit rectangles corresponding to extreme values of polygons associated with the one or more salient portions, and
• combining each minimal bounding fit rectangle to generate a union box;
• and
• cause a user interface to display a portion of the digital image that includes at least the one or more salient portions.

A non-transitory computer-readable storage medium having stored thereon a set of instructions, executable by at least one processor, for automatically determining and displaying salient portions of images, the instructions comprising:

• instructions for analyzing, by applying a saliency detection learning model, a digital image to generate at least one saliency map corresponding to at least one feature of interest in the digital image, the at least one saliency map selected from one or more saliency maps of the saliency detection learning model;
• instructions for determining one or more salient portions of the digital image at least by:
• generating a threshold image based on one or more edges of the at least one saliency map,
• determining, based on the threshold image, one or more minimal bounding fit rectangles corresponding to extreme values of polygons associated with the one or more salient portions, and
• combining each minimal bounding fit rectangle to generate a union box;
• and
• instructions for displaying, in a user interface, a portion of the digital image that includes at least the one or more salient portions.