Composite Image Generation Using Attribute Segmentation and Redaction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of creating a composite image of a suspect from multiple witness descriptions is time-consuming and difficult for untrained witnesses, as it requires selecting and combining various facial attributes accurately.
Innovation Solution
A method and apparatus that present witnesses with a series of attributes, allowing them to select and refine choices by analyzing and redacting images from a database of incident scene images, ensuring more accurate attribute selection and generating a composite image efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional composite image creation methods are used, then the process requires significant time and effort, but the accuracy of attribute selection can be maintained
Solution Approach 1:
The patent segments the composite image creation process into distinct attribute categories (hair, eyes, nose, mouth, etc.), allowing witnesses to select from pre-defined options for each category. This segmentation simplifies the complex task of creating a composite image by breaking it down into manageable, independent selection steps, thereby reducing time loss while maintaining selection accuracy.
Solution Approach 2:
The system performs preliminary actions by pre-processing and organizing attribute data from incident scene images before witness selection. Attributes are extracted, categorized, and prepared in advance, so when witnesses need to make selections, the work is already partially done. This preliminary preparation significantly reduces the time required during the actual witness interaction while ensuring accurate attribute representation.
2Reliability
If untrained witnesses are asked to create composite images, then the process becomes very difficult, but the reliability of the witness's memory can be preserved
Solution Approach 1:
The patent introduces an intermediary system that acts as a mediator between the witness's memory and the composite image creation process. The system provides structured attribute options based on incident scene data, translating the witness's verbal descriptions into visual selections. This intermediary layer protects witness memory reliability by guiding them through systematic choices while making the operation easy for untrained individuals.
Solution Approach 2:
The system enables witnesses to serve themselves by providing them with direct access to attribute selection interfaces that adapt to their responses. The automated system guides witnesses through the process, presenting relevant attributes and capturing their selections without requiring manual intervention from investigators. This self-service approach maintains reliability by letting witnesses make their own choices while simplifying the operation through automated guidance.
3Measurement precision
If more attribute options are provided to witnesses, then the accuracy of the composite image improves, but the complexity of the selection process increases
Solution Approach 1:
The patent implements dynamic attribute selection where the system adapts the number and type of attributes presented based on witness responses and incident context. The attribute set is not static but dynamically adjusted to provide relevant options while filtering out unnecessary complexity. This dynamic approach maintains high accuracy by providing comprehensive options when needed while reducing complexity through contextual filtering.
Solution Approach 2:
The system changes parameters of the attribute selection process based on incident scene data and witness progress. Attributes are organized hierarchically with different levels of detail, allowing the system to adjust the granularity of options presented. This parameter adjustment enables comprehensive attribute coverage for accuracy while managing complexity by only displaying relevant detail levels at each selection stage.
Data Source
AI summary
A method and apparatus for creating a composite image of a person is provided herein. During operation, a witness is presented with a plurality of attributes of an individual. A witness selection of an attribute is received from the presented attributes. Images of individuals who were at an incident scene at a time of the incident are received and analyzed to determine images of those individuals with attributes similar to the attribute selected by the witness. The images of those individuals with the attributes similar to the attribute selected by the witness are processed by redacting information from the images to produce a plurality of processed images. The witness is then presented with the plurality of processed images.