3D Body Data Generation from 2D Images via Probability Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for obtaining accurate three-dimensional body measurements require specialized and expensive equipment, limiting their accessibility for private individuals and patients who need to monitor changes in their body shape, especially with the rise of telemedicine and the clothing industry's need for accurate body dimensions.
Innovation Solution
A method and system that generate three-dimensional body data from two-dimensional images using image segmentation techniques, probability maps, and Gradient Descent optimization to compare with a database of known models, allowing for easy and cost-effective monitoring without dedicated imaging equipment, using devices like webcams or smartphones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional body scanners with high resolution depth sensors and fixed patterns of light are used, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses two-dimensional images as a simplified copy or representation of the three-dimensional body, rather than directly capturing 3D data with complex depth sensors. The system creates a 3D body model by processing and comparing 2D images against a database of 3D models, effectively copying the body's appearance from multiple angles to infer its three-dimensional structure without requiring complex imaging hardware.
Solution Approach 2:
The patent replaces the mechanical depth sensing system with an image processing system. Instead of using physical depth sensors, structured light patterns, or active cameras to directly measure distances, the system uses standard 2D images and applies computational algorithms (image segmentation, probability maps, Gradient Descent optimization) to extract three-dimensional information, substituting mechanical measurement with computational inference.
2Measurement precision
If conventional body scanners with complex equipment are used, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The system uses simple 2D images that can be captured by standard cameras instead of requiring users to operate complex 3D scanning equipment. Users can easily take photos with their smartphones or webcams, and the system automatically processes these simple images to generate accurate 3D body measurements, making the technology accessible to private individuals without specialized training or equipment.
Solution Approach 2:
The system performs automatic image processing, segmentation, and 3D model generation without requiring user intervention in complex operations. The automated comparison against the database of 3D models and the use of Gradient Descent optimization eliminate the need for users to manually adjust parameters or operate sophisticated hardware, allowing anyone to easily obtain accurate body measurements.
3Measurement precision
If dedicated imaging equipment is used, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system uses readily available 2D images that can be captured instantly with standard cameras, eliminating the need for time-consuming 3D scanning procedures. Users can quickly take photos and upload them to the system, which then automatically generates 3D body data through image processing and database comparison, significantly reducing the time required compared to conventional scanning methods.
Solution Approach 2:
The patent replaces time-consuming mechanical scanning processes with rapid 2D image capture and automated computational processing. Instead of requiring users to stand still for extended periods while complex scanners capture multiple depth measurements, the system processes standard images quickly using image segmentation and Gradient Descent optimization, achieving accurate 3D data generation in a fraction of the time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This invention relates to a method of generating three dimensional body data of a subject. The method includes the following steps. First or more source images of the subject are captured using a digital imaging device. The one or more source images are partitioned into a plurality of segments or probability distributions using one or more segmentation method, heuristics and/or predetermined mappings. The results of each segmentation method are combined to produce one or more unique probability maps representing the subject and the unique probability maps are then compared with a database of representations of three dimensional bodies to determine a closest match or best mapping between the or each unique probability map and a 'representation determined from the database. Three dimensional body data and'or measurements of the subject is then generated and may be. output based on the best mapping.