Abstract: The disclosed technique detects a foreground region accurately even with a neural network. A first region detection unit detects, by a DNN, a predetermined foreground region in an inputted image. A weak region identification unit detects, by a DNN, a weak region having a possibility that misdetection or non-detection occurs during the region detection on the inputted image using a DNN that has the same detection target as that of the DNN of the first region detection unit. A second region detection unit detects, by a method other than a neural network, a foreground region that is a detection target same as that of the first region detection unit in the weak region detected by the weak region identification unit within the inputted image read from an input device. An integration unit integrates detection results by the first region detection unit and by the second region detection unit.

Abstract: An information processing apparatus generates, from a plurality of captured images obtained by a plurality of cameras, a plurality of silhouette images representing respective regions of a plurality of objects appearing in the plurality of captured images, obtains associating information representing a silhouette image associated with each object by associating the plurality of silhouette images with each object of the plurality of objects, and generates, based on the associating information corresponding to an object designated from the plurality of objects and a designated virtual viewpoint, a virtual viewpoint image of the designated object.

Abstract: An image processing apparatus has one or more memories storing instructions; and one or more processors executing the instructions to: acquire a foreground object image, the foreground object image being an image viewing a foreground object from a virtual viewpoint and including no background; acquire a background image rendered using computer graphics, the background image being an image viewing a CG space from the virtual viewpoint and including background; generate, based on two-dimensional information on a shape of the foreground object and information on a light in the CG space, a shadow image indicating a shadow of the foreground object corresponding to the CG space; and generate a combined image by combining the foreground object image, the background image, and the shadow image into a single image.

Abstract: The disclosed technique detects a foreground region accurately even with a neural network. A first region detection unit detects, by a DNN, a predetermined foreground region in an inputted image. A weak region identification unit detects, by a DNN, a weak region having a possibility that misdetection or non-detection occurs during the region detection on the inputted image using a DNN that has the same detection target as that of the DNN of the first region detection unit. A second region detection unit detects, by a method other than a neural network, a foreground region that is a detection target same as that of the first region detection unit in the weak region detected by the weak region identification unit within the inputted image read from an input device. An integration unit integrates detection results by the first region detection unit and by the second region detection unit.