Vision-Based Touch Detection Using Temporal Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current touch detection methods rely on touch sensors, which are not applicable in mixed reality environments where real objects without sensors need to be interacted with, necessitating an improved vision-based touch detection technique.
Innovation Solution
A method involving the processing of raw touch data using temporal filters, such as high pass and low pass filters, based on detected touch types, to accurately identify touch events in physical environments without the need for touch sensors, utilizing depth images and probability analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If vision-based touch detection is used instead of touch sensors, then real objects without sensors can be interacted with in mixed reality environments, but the accuracy and reliability of touch detection is reduced
Solution Approach 1:
The system performs preliminary classification of touch types (click vs. drag) by analyzing movement patterns and velocity across multiple frames before applying the appropriate filter. This preliminary action enables the system to prepare the correct processing mode in advance, improving detection accuracy without requiring touch sensors on real objects.
Solution Approach 2:
The system dynamically switches between different temporal filters (high-pass for clicks, low-pass for drags) based on the detected touch type. This dynamic adaptation allows the system to optimize detection accuracy for different interaction patterns, resolving the contradiction between versatility and precision by adjusting the detection mechanism in real-time.
2Measurement precision
If temporal filters are applied to raw touch data, then touch detection accuracy is improved, but processing time and computational complexity increase
Solution Approach 1:
The processing pipeline is segmented into distinct stages: raw data acquisition, touch type classification, filter selection, and filtered data generation. By dividing the processing into discrete segments, the system can apply computational optimizations at each stage and avoid unnecessary processing, reducing overall processing time while maintaining accuracy.
Solution Approach 2:
The system applies temporal filters selectively based on the detected touch type rather than processing all data with the same level of complexity. For example, simple click detections may require less intensive filtering than complex drag patterns, allowing the system to use partial processing action appropriate to each case and reduce unnecessary computational overhead.
Data Source
AI summary
Detecting a touch includes determining, from first raw touch data, an estimated touch type for a first plurality of frames, obtaining second raw touch data comprising a probability that a touch has occurred for each of a second plurality of frames, applying the selected filter to the additional raw data to obtain processed data, and detecting whether a touch has occurred for a particular frame based on the processed data.