Probabilistic Palm Rejection via Spatiotemporal Touch Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current touch devices lack effective palm rejection features, leading to ergonomic issues and accidental inputs due to the inability to reliably distinguish between stylus and finger inputs, especially in the absence of special hardware, and existing software solutions are inadequate.
Innovation Solution
A novel, iterative, probabilistic approach that uses spatiotemporal touch features and decision trees to classify touch points as either stylus or palm inputs, without requiring initial configuration or specific screen orientation, leveraging features like touch area, velocity, and temporal relationships to improve classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If special hardware (ultrasonic transducers, infrared sensors, Hall Effect sensors) is used to distinguish stylus from finger input, then input discrimination accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces physical sensor hardware with a software-based capacitive touch analysis system. Instead of using ultrasonic transducers, infrared sensors, or Hall Effect sensors, the invention analyzes capacitive touch data already generated by the touchscreen controller, using machine learning algorithms to distinguish between stylus, finger, and palm inputs based on touch characteristics such as contact area, pressure distribution, and temporal patterns.
Solution Approach 2:
The patent creates a virtual model of touch inputs by analyzing capacitive characteristics. Rather than directly sensing physical properties with dedicated hardware, the system creates representations of touch events from capacitive data and uses computational methods to classify them, effectively copying the discrimination function from hardware sensors to software processing.
2Reliability
If special hardware sensors are added to the device, then stylus detection capability is improved, but manufacturing cost and device size increase
Solution Approach 1:
The invention substitutes dedicated stylus detection hardware with software processing of existing capacitive touch data. The system uses machine learning models trained on capacitive touch characteristics to reliably detect stylus inputs without requiring additional sensors, thereby maintaining stylus detection capability while eliminating the need for specialized hardware components that would increase manufacturing cost.
Solution Approach 2:
The patent makes the capacitive touchscreen serve multiple functions: it detects both finger and stylus inputs, and also provides palm rejection, using the same hardware infrastructure. The machine learning system analyzes capacitive data to distinguish between different input types, allowing the touchscreen to universally handle various input methods without requiring separate detection mechanisms for each.
3Area of stationary object
If geometric models are used to define palm rejection regions, then palm rejection coverage is improved, but flexibility and adaptability to different user inputs deteriorate
Solution Approach 1:
The patent transitions from static geometric palm rejection regions to dynamic, data-driven classification. Instead of using fixed geometric shapes to define rejection zones, the system continuously analyzes capacitive touch characteristics and temporal patterns to dynamically determine whether each touch point should be rejected, allowing the rejection behavior to adapt to different users, hand sizes, and input styles in real-time.
Solution Approach 2:
The invention changes the parameters used for palm rejection from fixed geometric coordinates to variable capacitive characteristics such as contact area, pressure distribution, and touch duration. The machine learning model adjusts classification decisions based on these parameters, enabling flexible adaptation to different palm sizes and input patterns without being constrained by predetermined geometric boundaries.
4Device complexity
If software-only solutions are used for palm rejection, then device complexity and cost are reduced, but classification accuracy and reliability worsen
Solution Approach 1:
The patent enhances software-based classification by utilizing multiple capacitive parameters including contact area, pressure distribution, touch duration, and temporal patterns. The machine learning model processes these parameters to achieve high classification accuracy, demonstrating that sophisticated software analysis of existing capacitive data can match or exceed the performance of simpler hardware-based solutions.
Solution Approach 2:
The invention introduces machine learning algorithms as an intermediary layer between the capacitive touchscreen and the input processing system. This intermediary analyzes capacitive touch characteristics and temporal patterns to make intelligent classification decisions, bridging the gap between simple capacitive sensing and accurate input discrimination without requiring additional hardware sensors.
Data Source
AI summary
The disclosed subject matter is a palm rejection technique utilizing temporal features, iterative classification, and probabilistic voting. Touch events are classified based on features periodically extracted from time windows of increasing size, always centered at the birth of the event. The classification process uses a series of decision trees acting on said features.