Face-Aware Image Synthesis for Video Call Continuity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for video calls require manual user intervention to select and synthesize images, disrupting the conversation flow and often fail to harmoniously integrate the synthesized images with the face within the image, as they are not adjusted for the face's location or size.
Innovation Solution
An image processing method and apparatus that automatically obtain an image with a face, determine if preset conditions are met, and synthesize additional images at the appropriate location and size based on face part location and size information, ensuring seamless integration and conversation continuity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual image selection is required during video call, then image synthesis can be performed, but conversation flow is lost
Solution Approach 1:
The system performs preliminary actions by automatically detecting face locations and preparing synthesis images before the user needs to send any message. The face detection and image preparation occur in advance based on the video feed, so when the user initiates a synthesis request, the processing is already ready, eliminating the need for manual selection during conversation.
Solution Approach 2:
The system serves itself by automatically detecting faces and preparing synthesis images without requiring user intervention. The automatic face detection and image synthesis preparation occur based on the video feed alone, making the system self-sufficient and eliminating the need for manual image selection by the user during the conversation.
2Ease of manufacture
If image synthesis is performed at predefined location, then synthesis process is simple, but synthesized image is not harmonized with face
Solution Approach 1:
The system applies local quality by detecting the specific location of the face within the image and using that information to determine where to place the synthesized image. Instead of using a fixed predefined location, the system adapts the synthesis location to match the actual face position, ensuring the synthesized image is harmoniously integrated with the detected face while maintaining a relatively simple synthesis process.
Data Source
AI summary
A method of processing an image is provided. The method includes obtaining an image including a face, obtaining information about whether a preset condition is satisfied after the image is obtained, obtaining location information of a face part in the image when the preset condition is satisfied, and obtaining a synthesis image by adding an image corresponding to the satisfied condition at a location of the face part to the image.