Composite Selfie Image Normalization for Remote Users
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for composite selfie image composition result in artificial-looking images with unpleasant artifacts, failing to effectively merge self-portraits of remote users into a natural and cohesive image.
Innovation Solution
An image processing method and device that normalize face sizes and positions, overlay self-portraits on a selected background, and adjust facial expressions to create a cohesive composite selfie image, using image-capture devices and a server for communication and background removal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If existing image processing techniques are used to merge self-portraits, then the composite image can be created, but the result looks artificial and contains unpleasant artifacts
Solution Approach 1:
The system performs preliminary actions by detecting faces, removing backgrounds, and normalizing face sizes and positions before the final composition step. This preparatory processing ensures that when images are merged, they align properly and appear natural, avoiding artifacts while maintaining ease of composite image creation
Solution Approach 2:
The patent introduces an intermediary processing stage that acts as a mediator between raw self-portraits and the final composite image. This intermediary step includes background removal, face detection, and normalization processes that transform individual portraits into standardized components ready for seamless integration, thereby improving visual quality without complicating the overall process
2Adaptability or versatility
If multiple self-portraits are merged into one image, then remote users can be included in the selfie, but the facial expressions and positions become misaligned and unnatural
Solution Approach 1:
The system applies local quality adjustments by detecting specific facial features in each self-portrait and applying targeted transformations to individual face regions. Face normalization, positioning, and expression adjustments are applied locally to each detected face rather than globally to the entire image, ensuring that each user's portrait is optimally aligned while maintaining the ability to include multiple remote users
Solution Approach 2:
The patent utilizes parameter changes by systematically adjusting key parameters of each self-portrait including face size, position coordinates, orientation angles, and expression characteristics. These parameter transformations are applied based on detected facial features to ensure all users appear naturally aligned in the composite image while maintaining the versatility to include remote participants
Data Source
AI summary
A device and method for composite selfie image composition for remote users, includes a first image-capture device that communicates a first notification message to a second image-capture device associated with a second user. The first notification message includes a request to capture a second self-portrait image of the second user, a user-identifier of a first user associated with the first image-capture device and first face coordinates of the first user. The second self-portrait image and second face coordinates of the second user is received from the second image-capture device. By face normalization on a portion of the second self-portrait image, a second face size of the second user is adjusted in accordance with a first face size in a first self-portrait image of the first user. A first output image, for example, a composite selfie, is generated based on at least the face normalization.