Face Image Grouping Using Product Statistics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional face detection methods require prior training with known faces and supervision, making them inefficient for accurately grouping face images in image products, especially when dealing with unknown individuals.
Innovation Solution
A computer-implemented method that uses probabilistic criteria and similarity functions to automatically categorize face images without prior training, employing non-negative matrix factorization to group faces into modified groups based on image-product statistics, allowing for the creation of personalized image products.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional face detection methods are used, then face recognition can be performed, but prior training with known faces and supervision is required which reduces efficiency
Solution Approach 1:
The system performs self-service by automatically grouping face images without requiring external training data or supervision. The face grouping module autonomously analyzes facial features and clusters images by identity, eliminating the need for manual training with known faces while maintaining high grouping accuracy
Solution Approach 2:
The system performs preliminary face detection and feature extraction before grouping, preparing face images and their characteristics in advance. This preliminary processing enables the subsequent unsupervised grouping to proceed efficiently without requiring training at the grouping stage
2Ease of operation
If manual design steps are used for image products, then customization is possible, but time consumption increases significantly
Solution Approach 1:
The system automatically performs image selection, layout generation, and face grouping without requiring manual user intervention for each design step. The automated image product generation module creates personalized products by autonomously selecting relevant images and arranging them according to detected face relationships, dramatically reducing design time while preserving customization through algorithmic intelligence
Solution Approach 2:
The system performs preliminary face grouping and image categorization before the actual product design process. By pre-organizing images according to detected facial identities and relationships, the system prepares structured data that enables rapid automated layout generation, eliminating the need for manual image selection and arrangement
3Measurement precision
If face images are grouped without accurate methods, then processing is faster, but grouping accuracy decreases leading to incorrect person identification
Solution Approach 1:
The system segments the face grouping process into distinct stages: face detection, feature extraction, similarity calculation, and clustering. Each stage processes specific aspects independently, allowing optimized computation at each step while maintaining overall accuracy through the coordinated sequence of operations
Solution Approach 2:
The system transforms face images into high-dimensional feature vectors and performs grouping in this abstract feature space rather than direct pixel space. This dimensional transformation enables more accurate similarity measurement and clustering while reducing computational complexity through dimensionality reduction techniques
Data Source
AI summary
A computer-implemented method for creating an image product by accurately grouping faces includes receiving an initial set of face groups for a plurality of face images, training classifiers between pairs of face groups in the initial set of face groups using image-product statistics by a computer processor, classifying the plurality of face images by classifiers to output binary vectors for the plurality of face images by the computer processor, calculating a value for an improved similarity function using the binary vectors for each pair of the plurality of face images, grouping the plurality of face images into modified face groups based on values of the binary similarity functions by the computer processor, and creating an image product based at least in part on the modified face groups.