Face Recognition System for Context-Aware Image Description
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic textual description systems for images fail to capture the relationships between individuals in images, limiting user understanding of interrelations among persons or objects.
Innovation Solution
A method and apparatus that recognize faces in images, determine personal identities, and generate descriptions using photographing information, employing geometrical and model-based methods, and natural language processing to create accurate and context-rich descriptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If simple descriptions are provided separately for persons or objects in images, then the system is easy to operate and process, but the user cannot obtain the interrelation among the persons
Solution Approach 1:
The patent merges separate person descriptions into a unified description that includes relational information. Instead of providing independent descriptions for each person, the system generates a combined description that captures the relationships between multiple persons in the image, thereby reducing information loss while maintaining manageable processing complexity through integrated analysis.
Solution Approach 2:
The description generation module performs multiple functions simultaneously: it identifies persons, determines their relationships, and generates comprehensive descriptions all within a single processing pipeline. This multi-functional approach allows the system to provide rich interrelation information without requiring separate specialized systems for each function.
2Measurement precision
If face recognition and personal identity determination are performed, then accurate and detailed information about individuals can be provided, but the processing time and computational resources increase
Solution Approach 1:
The system performs face recognition and personal identity determination as preliminary steps before generating the final description. By completing these identification tasks in advance, the system ensures accurate person information is available when needed, though the processing time is managed through efficient sequential execution rather than parallelization.
Solution Approach 2:
The system uses feedback from face recognition results to refine the personal identity determination and subsequent description generation. This iterative feedback mechanism ensures high accuracy in person identification while optimizing processing efficiency by adjusting the depth of analysis based on recognition confidence levels.
Data Source
Figure 1~2
Figure 3~5
Figure 6~7
AI summary
The present invention is related to a picture processing method and apparatus, a computer program and a recording medium. The method comprises: recognizing a face in a picture of a user; determining a personal identity corresponding to the recognized face, the personal identity comprising at least one of an identification of a person corresponding to the face and a relationship between the person corresponding to the face and the user; acquiring photographing information of the picture, the photographing information comprising at least one of a photographing time and a photographing location of the picture; and generating a description of the picture according to the personal identity and the photographing information. With this technical solution, a description of the picture can be generated more accurately, the intelligence for automatic description of a picture is improved to be closer to human's ability to describe the picture, the user can quickly get an accurate knowledge of every picture, and user experience is improved.