Facial Region Extraction for Small Device Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems lack an efficient method for automatically processing and selecting facial images from large catalogs, particularly for small devices where full-body or group images may render celebrities unrecognizable.
Innovation Solution
A system and method for automatically processing images to extract and crop facial regions, optionally applying masking and expanding images to improve recognition and reduce data transmission requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full-body or group images are displayed on small devices, then the complete subject and context are preserved, but the celebrity becomes unrecognizable
Solution Approach 1:
The patent extracts the facial region from the original image by detecting face boundaries and cropping only that portion. This extraction allows the system to present a focused view of the celebrity's face on small devices, ensuring recognizability while eliminating unnecessary background information from full-body or group shots.
Solution Approach 2:
The patent segments the image into relevant facial region and irrelevant background regions. By identifying and isolating the face based on detected boundaries, the system separates the critical information (face) from non-critical information (body, background), enabling optimized display on limited-screen devices.
2Reliability
If automatic face extraction and cropping is performed, then face recognition performance is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary face detection and boundary identification before the actual cropping operation. By pre-determining the facial region through detection algorithms and storing boundary information, the system prepares the necessary parameters in advance, enabling faster subsequent cropping operations without repeating the detection process.
Solution Approach 2:
The patent creates a cropped copy of only the facial region from the original image. This copying operation transfers only the essential face information to the output, reducing the amount of data that needs to be processed further while maintaining face identification accuracy, thereby reducing overall processing time.
3Measurement precision
If large image catalogs are processed manually, then quality control and selection accuracy are improved, but processing efficiency and scalability deteriorate
Solution Approach 1:
The patent implements self-service automation where the system independently detects faces, determines boundaries, crops images, and ranks results without requiring manual intervention. The automated detection and processing algorithms perform quality control and selection tasks that would otherwise require human operators, enabling high-volume processing of large catalogs at scale.
Solution Approach 2:
The patent replaces manual mechanical image processing with automated computational algorithms. Instead of human operators visually inspecting and selecting images, the system uses computer vision algorithms to detect faces, determine boundaries, and automatically crop and rank images, dramatically increasing processing speed and scalability while maintaining selection accuracy.
Data Source
AI summary
Various embodiments contemplate systems, architectures and methods for extracting and selecting headshots of human or non-human entities from catalogs of images of such subjects. The methods described may find and extract faces from within groups of subjects, verify that the extracted faces correspond to the desired subject, determine cropping or masking regions, or both, of rectangular, circular, elliptical or some other geometry to provide an easily recognized image of the desired subject, expand the output image by synthesizing pixels as may be needed for a desired cropping or masking region, select preferred images among a collection of images of the desired subject, and perform other useful functions. The resulting output images may be in either direct form, reference form or both forms.