3D Human Body Estimation via Spatial Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human body estimation technologies in augmented reality struggle to accurately represent the 3D shape of a person, particularly in interactions with virtual objects, as they fail to account for spatial relationships and nuances like the space between body parts, leading to unrealistic representations.
Innovation Solution
A method and apparatus that receive a 2D image frame from a camera, determine 2D appearance parameters using a convolutional neural network, and estimate 3D human body models by combining camera and spatial parameters, including rotation and translation, to accurately project 3D human bodies, enabling realistic interactions with virtual objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing human body estimation technologies are used, then the processing speed is fast, but the accuracy of 3D shape representation is poor
Solution Approach 1:
The estimation system is divided into two independent modules: a 2D appearance parameter determination module that analyzes 2D images to extract appearance features, and a 3D appearance parameter determination module that uses these 2D parameters along with spatial information to estimate 3D shape. This segmentation allows each module to specialize in specific tasks, improving overall 3D shape representation accuracy while keeping individual module complexity manageable.
Solution Approach 2:
The system transitions from 2D image analysis to 3D shape estimation by introducing spatial dimension parameters (rotation and translation) that represent the relative relationship between the camera and the human body. This dimensional transformation enables accurate 3D shape representation by incorporating depth and spatial orientation information that is absent in 2D images alone.
2Measurement precision
If 2D appearance parameters are used alone, then the processing is simple, but the 3D shape estimation is inaccurate
Solution Approach 1:
The system merges 2D appearance parameters (extracted from image analysis) with 3D space parameters (rotation and translation representing camera-human body relative relationships) to determine 3D appearance parameters. This combination allows the system to leverage simple 2D analysis while incorporating spatial information, achieving accurate 3D shape estimation without excessive processing complexity.
Solution Approach 2:
The 2D appearance parameters serve as an intermediary that bridges 2D image data and 3D shape estimation. Instead of directly estimating 3D shape from raw images, the system first extracts 2D appearance parameters as intermediate features, which are then combined with spatial parameters to produce accurate 3D appearance parameters, simplifying the overall processing pipeline.
3Reliability
If spatial parameters are not considered, then the estimation process is simpler, but the representation of body part relationships is unrealistic
Solution Approach 1:
The system performs preliminary determination of 3D space parameters (rotation and translation representing camera-human body relative relationships) before finalizing the 3D appearance parameter estimation. By pre-establishing the spatial relationship framework, the system ensures accurate body part spatial relationships in the final estimation without adding excessive complexity to the core estimation process.
Data Source
AI summary
A human body estimating method includes: receiving an image frame including a two-dimensional (2D) human body from a camera; determining a 2D appearance parameter of the 2D human body by analyzing the image frame; determining a three-dimensional (3D) appearance parameter of the 2D human body based on the 2D appearance parameter, a camera parameter of the camera, a 3D space parameter of the 2D human body; and estimating a 3D human body corresponding to the 2D human body based on the 3D space parameter and the 3D appearance parameter.