Image processing apparatus and control method for same

A control method for an image processing apparatus, the method comprising:

• obtaining a captured image;
• obtaining a likelihood map indicating a likelihood of existence of an object at each position of the captured image;
• obtaining, at each position of the captured image, a region tensor indicating a position and a size of an object with respect to each position;
• accepting first position coordinates with respect to the captured image;
• determining an object region corresponding to the first position coordinates, based on the region tensor and likelihoods of the likelihood map corresponding to two or more region candidates each indicated by the region tensor based on the first position coordinate;
• determining a first object region based on the first position coordinates and a second object region based on second position coordinates determined from the first object region;
• deciding a larger one of a first likelihood value corresponding to the first position coordinates in the likelihood map and a second likelihood value corresponding to the second position coordinates with respect to center of the object region in the likelihood map;
• and
• outputting the first object region to an external apparatus if it is decided that the first likelihood value is larger than the second likelihood value, and outputting the second object region to the external apparatus if it is decided that the first likelihood value is larger than the second likelihood value.

A non-transitory computer-readable recording medium storing a program for causing a computer to execute as an image processing apparatus, wherein the program causes the computer to:

• obtain a captured image;
• obtain a likelihood map indicating a likelihood of existence of an object at each position of the captured image;
• obtain, at each position of the captured image, a region tensor indicating a position and a size of an object with respect to each position;
• accept first position coordinates with respect to the captured image;
• determine an object region corresponding to the first position coordinates, based on the region tensor and likelihoods of the likelihood map corresponding to two or more region candidates each indicated by the region tensor based on the first position coordinate;
• determine a first object region based on the first position coordinates and a second object region based on second position coordinates determined from the first object region;
• decide a larger one of a first likelihood value corresponding to the first position coordinates in the likelihood map and a second likelihood value corresponding to the second position coordinates with respect to center of the object region in the likelihood map;
• and
• output the first object region to an external apparatus if it is decided that the first likelihood value is larger than the second likelihood value, and output the second object region to the external apparatus if it is decided that the first likelihood value is larger than the second likelihood value.