Motion Prediction Using Single Camera and Synthetic Occlusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion capture systems are expensive, cumbersome, and prone to tracking inaccuracies due to the need for multiple cameras and specific backgrounds, which can lead to poor recognition of limbs and interference from other objects.
Innovation Solution
A method and device for motion prediction using a single camera or 3D software to capture multi-view images, synthesize motion capture data, project a masking object to generate occluded limb images, and train a predictive model using deep neural networks or AI to predict joint information, allowing for accurate motion capture without specific backgrounds or clear fields.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple cameras are used to capture motion from different angles, then motion capture accuracy is improved, but system cost and complexity increase
Solution Approach 1:
The patent uses a single camera to capture motion and then generates synthetic multi-view images through image processing and 3D reconstruction algorithms. Instead of physically deploying multiple cameras, the system creates virtual copies of the motion data from different perspectives through computational methods, thereby achieving multi-view motion capture accuracy with single-camera simplicity
Solution Approach 2:
The patent replaces the mechanical system of multiple physical cameras with a computational system using a single camera combined with image processing algorithms. The mechanical complexity of multi-camera setup is substituted by software-based 3D reconstruction and synthetic view generation, maintaining measurement precision while reducing device complexity
2Measurement precision
If markers or sensors are worn by the subject, then motion capture accuracy is improved, but ease of operation deteriorates due to cumbersome equipment
Solution Approach 1:
The patent extracts the motion capture function from the subject by using a single external camera to observe and track the subject's movements. Instead of attaching markers or sensors to the subject's body, the system captures motion information from the outside through image processing, eliminating the need for subject wearables and improving ease of operation while maintaining tracking accuracy
Solution Approach 2:
The patent enables the subject to perform natural movements without any assistance from markers or sensors. The single camera system automatically tracks and captures motion data through image processing algorithms that identify and follow the subject's body parts, allowing the subject to move freely without being serviced by additional equipment
3Measurement precision
If a clear field or green screen background is used, then motion recognition accuracy is improved, but adaptability deteriorates due to background constraints
Solution Approach 1:
The patent creates a universal motion capture system that can operate in any background environment. The single camera combined with image processing algorithms can distinguish the subject from any background, making the system adaptable to diverse environments without requiring specific background conditions. The system performs multiple functions: capturing motion, processing images, and adapting to different backgrounds using the same hardware setup
Solution Approach 2:
The patent changes the approach from controlling background parameters (requiring green screens or clear fields) to processing image parameters through algorithms. The system adjusts image processing parameters to separate the subject from various backgrounds, enabling motion capture accuracy without background constraints and improving adaptability to different environments
4Adaptability or versatility
If other objects are present in the field, then system versatility is improved, but measurement precision deteriorates due to tracking errors
Solution Approach 1:
The patent introduces image processing algorithms as an intermediary between the single camera and the motion capture output. These algorithms act as a mediator that can distinguish the subject from other objects in the environment, filtering out interference from background elements and maintaining tracking accuracy even when other objects are present, thereby preserving measurement precision while allowing environmental versatility
Data Source
AI summary
A method for motion prediction is provided. The method includes: capturing, by a single camera, a plurality of cameras or 3D software, images of an object at multiple angles to generate multi-view images of the object; synthesizing motion capture (MoCap) data according to the multi-view images; projecting a masking object onto the object to generate multi-view training images, wherein the multi-view training images are images in which parts of limbs of the object are unoccluded and other parts of the limbs of the object are occluded; and using the motion capture data and the multi-view training images to train a predictive model.