GENERATING A GROUP PHOTO THAT INCLUDES A PHOTOGRAPHER

Smart overview of the Invention

A method and system are described for creating a composite image that includes the photographer in group photos. This is achieved without requiring support structures or remote controls. The process involves using a user device to capture a second image that complements a first image, ensuring the inclusion of all desired subjects, including the photographer.

Process

The method begins with a user device receiving a request to generate a composite image. A first image, containing one or more subjects, is analyzed to determine the device's pose during capture. The subjects are segmented from this image, and overlays are generated to guide the user in capturing a second image. This guidance is based on comparing the device's current pose to its previous pose.

Depth and Order

The system determines the depth of subjects in both images to aid in generating accurate overlays and composite images. It uses these depths and possibly a machine-learning model to decide the order of subjects in the final composite image, ensuring logical placement, such as having foreground subjects in front.

Guidance and Segmentation

The user is guided to adjust the device's position to match the previous pose, ensuring alignment of images. Before creating overlays, the system segments backgrounds and foregrounds to confirm the subjects' positions. This segmentation supports the accurate generation of the composite image.

Composite Image Generation

The final composite image is created by combining segmented subjects from both images. The system evaluates stitching scores to choose the best composite image. If gaps are detected, inpainting techniques fill them, ensuring a seamless and complete group photo.