Textured 3D Image Generation via Angular Orientation Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SLAM techniques face challenges in accurately generating 3D models from 2D images of varying quality and with offset facial features, such as blurry images and differing eye orientations, which affect the accuracy and consistency of 3D representation.
Innovation Solution
A method that processes multiple 2D images to determine the 3D angular orientation of an object, ranks images based on angular orientation, and selects images with aligned features to generate a textured 3D representation, ensuring that features like eyes are open and aligned, using cross-product calculations for orientation determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SLAM techniques are used to generate 3D models from multiple 2D images, then 3D representation can be created, but the accuracy and consistency deteriorate when images have varying quality and offset facial features
Solution Approach 1:
The system performs preliminary actions by determining 3D angular orientation of the object and ranking images based on this orientation before texture generation. This preliminary sorting and selection process ensures that the most suitable images (with proper feature alignment and quality) are chosen for texturing, thereby improving 3D model accuracy while adapting to varying image qualities
Solution Approach 2:
The system applies local quality by selecting specific images based on their suitability for different regions. It identifies images where facial features (eyes, nose, mouth) are properly aligned and visible, and uses those specific images for texturing corresponding regions of the 3D model, rather than uniformly using all images
2Quantity of substance
If multiple 2D images are processed to create 3D representation, then more data is available for modeling, but feature alignment and consistency become more difficult
Solution Approach 1:
The system changes parameters by determining 3D angular orientation from 2D image coordinates and using this orientation parameter to rank and select images. This parameter transformation and selection process maintains feature alignment consistency across multiple input images by choosing only those images that meet specific orientation and quality criteria
Solution Approach 2:
Before generating the texture, the system performs preliminary actions including determining angular orientation, ranking images, and selecting the most suitable images. This preliminary filtering ensures that only images with proper feature alignment are used, maintaining consistency even when processing multiple images
3Adaptability or versatility
If images with different angular orientations are used, then more viewing angles are captured, but the reliability of feature positioning deteriorates
Solution Approach 1:
The system applies local quality by evaluating each image's angular orientation and feature visibility locally. It selects images where specific features (eyes, nose, mouth) are properly positioned and visible from appropriate angles, ensuring reliable feature positioning for each selected image while maintaining overall viewing angle coverage
Solution Approach 2:
The system changes the orientation parameter by determining 3D angular orientation from 2D coordinates and using this to filter and rank images. This parameter-based selection ensures that only images with reliable feature positioning at various angles are used, balancing viewing angle coverage with positioning accuracy
Data Source
AI summary
A method for creating a 3D image is described. The method includes receiving a plurality of 2D images of an object, identifying respective 2D positions of one or more features in each of the plurality of 2D images, generating a 3D representation of the object, based on the respective 2D positions of the one or more features in each of the plurality of 2D images, and determining a 3D angular orientation of the 3D representation of the object based on relative positions of the one or more features in the 3D representation of the object. Related systems, devices and computer program products are also described.