Room Axis Estimation Using Majority Voting in AR Scans
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately estimating the axis of a room during a computer vision scan due to rotated objects, leading to user confusion and poor capture quality.
Innovation Solution
A computer-implemented method using a majority voting process to determine the room axis based on local axis orientations of walls and objects, utilizing augmented reality feedback to align the scan with the natural axis of the computer screen.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional scanning methods are used to capture room images, then the scanning process can be performed, but the room axis alignment with screen axes becomes inaccurate due to rotated objects
Solution Approach 1:
The patent segments the axis estimation problem into multiple local axis estimations from different objects (walls, furniture, etc.) in the room. Each object's local axis is estimated independently, then these local estimates are aggregated through a voting process to determine the final room axis. This segmentation allows the system to handle rotated objects by treating each object's orientation separately rather than requiring global alignment.
Solution Approach 2:
The patent merges multiple local axis orientation estimates from different objects and features in the room through a voting process. Each detected object (wall, furniture, etc.) contributes its local axis information, and the system combines these contributions to determine the dominant room axis orientation. This merging approach robustly handles the presence of rotated objects by aggregating evidence from multiple sources.
2Measurement precision
If users manually adjust capture techniques to correct alignment issues, then some alignment improvement may be achieved, but user confusion and frustration increase
Solution Approach 1:
The patent implements self-service by enabling the system to automatically estimate and correct room axis alignment without requiring user intervention. The computer vision system independently analyzes captured images, estimates local axes from various objects, performs voting to determine the room axis, and applies the necessary rotation correction automatically. This eliminates the need for users to manually adjust capture techniques or understand alignment concepts.
Solution Approach 2:
The patent employs feedback by using the estimated room axis orientation to guide the rendering and display of the 3D model. The system calculates the angle between the estimated room axis and the screen's natural axis, then uses this information to rotate the rendered view accordingly. This feedback loop ensures that the final presentation is always properly aligned with the screen coordinates, providing users with correctly oriented output without manual adjustment.
3Productivity
If multiple objects with different orientations are present in the room, then the room can be fully captured, but determining a consistent room axis becomes difficult
Solution Approach 1:
The patent applies partial action by estimating local axes from only the most prominent and reliably detectable objects in the room, rather than requiring analysis of every object. The voting process aggregates contributions from a sufficient subset of objects to determine the dominant axis orientation, without needing to process all objects equally. This approach maintains capture completeness while achieving accurate axis determination through selective focus on key features.
Data Source
AI summary
Systems and methods are provided for estimating an axis of a room based on a computer vision scan of the room and its contents, and a process of axis majority voting. A method is provided that includes capturing a plurality of images of the interior environment having one or more walls and a plurality of objects, each wall and object having a surface oriented along a corresponding plane; determining, based on information received from an AR engine, local axis orientations for at least a portion of the walls and objects. The method includes estimating a room axis of the interior environment based on a voting process, wherein a majority of matching local axis orientations are utilized to determine the room axis; and outputting an indication of the room axis.