Depth Hull 3D Model Generation for E-Commerce
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in purchasing items online due to the inability to view or touch products before buying, leading to hesitance and potential dissatisfaction with the actual item's size or style when received, as online photographs may not accurately represent the product.
Innovation Solution
A low-resource three-dimensional model generation technique that uses a combination of multi-view stereo and silhouette mask data to create a photorealistic representation of objects, effectively handling various surface types, including reflective and dark surfaces, and concave elements, by generating a dense point cloud and silhouette masks from multiple viewpoints and using a depth hull reconstruction method to generate a surface mesh.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional online photograph viewing is used, then purchasing convenience is improved, but product representation accuracy deteriorates
Solution Approach 1:
The patent creates accurate three-dimensional copies of physical objects by capturing images from multiple viewpoints and reconstructing the object's geometry. This digital replica preserves the true shape, size, and surface characteristics of the object, allowing customers to view and interact with a faithful representation online before purchasing.
Solution Approach 2:
The patent transitions from two-dimensional photographs to three-dimensional representations by capturing images from multiple angles and viewpoints. This dimensional enhancement provides customers with comprehensive spatial understanding of the object's shape, size, and features, resolving the limitation of flat images in accurately representing three-dimensional products.
2Measurement precision
If multiple viewpoint images are captured and processed, then three-dimensional model accuracy is improved, but computational resource requirements worsen
Solution Approach 1:
The patent employs a depth hull reconstruction method that processes a subset of viewpoint information rather than all possible views. By selectively using silhouette masks and depth data from multiple viewpoints to construct a bounding volume, the system achieves accurate three-dimensional reconstruction while reducing the computational burden of processing every possible image detail.
Solution Approach 2:
The patent divides the complex task of three-dimensional reconstruction into separate processing stages: capturing images from multiple viewpoints, generating silhouette masks, creating depth maps, and constructing the depth hull. This segmentation allows each processing step to be optimized independently, reducing overall computational resource requirements while maintaining reconstruction accuracy.
Data Source
AI summary
Approaches described and suggested herein relate to generating three-dimensional representations of objects to be used to render virtual reality and augmented reality effects on personal devices such as smartphones and personal computers, for example. An initial surface mesh of an object is obtained. A plurality of silhouette masks of the object taken from a plurality of viewpoints is also obtained. A plurality of depth maps are generated from the initial surface mesh. Specifically, the plurality of depth maps are taken from the same plurality of viewpoints from which the silhouette images are taken. A volume including the object is discretized into a plurality of voxels. Each voxel is then determined to be either inside the object or outside of the object based on the silhouette masks and the depth data. A final mesh is then generated from the voxels that are determined to be inside the object.