Image Recovery Using Active Appearance Models for Missing Face Elements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately recover missing elements in images, especially unique-shaped features like eyes, noses, or mouths, due to lack of identical components in surrounding regions, and fail to correct images with specific color signal loss or blurriness.
Innovation Solution
An image processing method using Active Appearance Models (AAM) that applies a statistical model of a predetermined structure to an input image to recover missing elements by correlating the structure with a model representing the structure, including principal component analysis for shape and luminance representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional interpolation methods are used to remove defective areas, then simple regions like sky or ocean can be corrected, but unique-shaped features like eyes, noses, or mouths cannot be recovered
Solution Approach 1:
The patent creates a mathematical model by copying and analyzing the statistical characteristics of complete structures from multiple reference images. This model captures the typical relationships between different components (e.g., eyes, nose, mouth) and uses this copied knowledge to reconstruct missing or defective unique features in the target image, rather than relying on local interpolation.
Solution Approach 2:
The patent transforms the image reconstruction problem into a parameter optimization problem. By representing the image structure through parameters (such as shape coefficients and appearance coefficients in AAM), the system can systematically adjust these parameters to find the most likely complete structure, enabling recovery of unique features that conventional pixel-level interpolation cannot handle.
2Measurement precision
If statistical modeling is applied to recover unique structure elements, then accurate recovery of missing face elements is achieved, but the processing complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-building a mathematical model from a database of reference images before the actual restoration task. This model captures the statistical relationships of structures in advance, so when restoration is needed, the system only needs to apply the pre-computed model to the target image, significantly reducing the computational complexity during the actual restoration process while maintaining high accuracy.
3Ease of operation
If existing correction methods are used, then user operation is simple, but the methods cannot correct images with specific color signal loss or blurriness
Solution Approach 1:
The patent implements self-service by enabling the system to automatically detect defective areas, select appropriate reference images, and perform restoration without requiring user intervention. The mathematical model autonomously identifies missing elements and reconstructs them by leveraging statistical relationships from the reference database, maintaining ease of operation while significantly improving correction effectiveness for various defect types including color signal loss and blurriness.
Data Source
AI summary
A face portion of an input image, an example of a predetermined structure, is applied to a mathematical model by the image recovery/addition section to recover a missing element of the face portion in the input image. The mathematical model is generated by a predetermined statistical method, such as the AAM scheme or the like, based on a plurality of sample images representing the face portion including the recovery target element. Thereafter, the face portion is reconstructed to include the missing element based on the parameter corresponding to the face portion obtained by applying the face portion to the model, and the face portion of the input image is replaced by the reconstructed face portion to produce a restored image by the image reconstruction section.