3D Gesture Recognition Using Singular Points of Interest
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human-to-computer interfaces struggle with accurately detecting and tracking 3D gestures performed by a single hand, particularly in varying distances and lighting conditions, and fail to provide robust recognition of simultaneous gestures without false positives, limiting natural interaction capabilities.
Innovation Solution
A method using a 3D range finding imaging system to detect singular points of interest on a hand, such as finger tips and the palm center, allowing for robust detection and tracking of gestures like pointing, pinching, and activation, even when performed simultaneously, by employing a combination of point cloud analysis and computational means for accurate gesture recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional 2D cameras are used for detecting fingers, then the system can capture light in the visible spectrum, but finger-like objects may be incorrectly identified and tracking may be lost due to dependency on scene illumination
Solution Approach 1:
The patent replaces conventional 2D optical cameras with a 3D range finding imaging system that actively emits light and measures time of flight. This substitution eliminates dependency on ambient scene illumination by using its own light source and active ranging mechanism, thereby improving gesture detection reliability in varying lighting conditions.
Solution Approach 2:
The patent changes the detection parameter from 2D image intensity (affected by illumination) to 3D time of flight measurements. By measuring the time for light to travel to and from the target, the system obtains depth information that is independent of ambient lighting conditions, resolving the illumination dependency problem.
2Measurement precision
If colour information is used to detect hand parameters, then the system can identify palm centre and hand extremities, but the method cannot operate in dark environments where colours may not be distinguished
Solution Approach 1:
The patent replaces passive colour-based detection with active 3D time of flight imaging. Instead of relying on colour information that requires ambient light, the system uses active light emission and time measurement to detect hand parameters, enabling operation in dark environments while maintaining detection precision.
Solution Approach 2:
The patent transitions from 2D colour-based hand parameter detection to 3D spatial detection using time of flight measurements. By adding the depth dimension and using active ranging, the system achieves environmental adaptability across varying lighting conditions while maintaining precise hand parameter detection.
3Ease of operation
If relative distance measurements are used for gesture detection, then the system can assess hand openness, but the measurements cannot be used to point and grab virtual objects accurately at various distances
Solution Approach 1:
The patent replaces relative distance measurement systems with absolute 3D time of flight ranging. The active light emission and time measurement provide true absolute distance values in metric units, enabling accurate pointing and grabbing of virtual objects at various distances while maintaining ease of gesture operation.
4Productivity
If a single hand performs multiple simultaneous gestures, then the interaction efficiency increases, but the system produces false positives and cannot reliably distinguish simultaneous gestures
Solution Approach 1:
The patent segments the hand into multiple tracked points of interest (fingertips, palm center, hand extremities) and independently tracks their 3D positions over time. This segmentation enables the system to distinguish multiple simultaneous gestures by analyzing the independent motion patterns of different hand segments, reducing false positives while maintaining high interaction efficiency.
Solution Approach 2:
The patent implements continuous feedback by tracking the temporal evolution of multiple hand points in 3D space. The system analyzes the coherent motion patterns of segmented hand points over time, providing feedback that distinguishes intentional simultaneous gestures from false positives through consistent spatial-temporal relationships.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables reliable and intuitive 3D gesture-based interactions with a computer system, providing accurate feedback and allowing for simultaneous pointing and activation gestures without false detection, enhancing the ergonomics and efficiency of human-computer interactions.
Implementation Method 1
A method using a 3D range finding imaging system to detect singular points of interest on a hand
Data Source
AI summary
Described herein is a method for enabling human-to-computer three-dimensional hand gesture-based natural interactions from depth images provided by a range finding imaging system. The method enables recognition of simultaneous gestures from detection, tracking and analysis of singular points of interests on a single hand of a user and provides contextual feedback information to the user. The singular points of interest of the hand: include hand tip(s), fingertip(s), palm center and center of mass of the hand, and are used for defining at least one representation of a pointer. The point(s) of interest is/are tracked over time and are analyzed to enable the determination of sequential and/or simultaneous “pointing” and “activation” gestures performed by a single hand.