Customizable User Input Recognition for Accessibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing recognition systems, such as voice and facial recognition, have limited customizability to individual users, particularly for children with complex communication needs or cerebral palsy, as they have non-standardized speech and movement patterns that require adaptable communication technologies to effectively convey intentions.
Innovation Solution
A customizable recognition system that uses a processor to differentiate and recognize individual speech and gesture patterns through a trained classifier linked to a user identifier, allowing for the execution of specific commands on computer devices, such as laptops or tablets, using audio/video inputs like tongue gestures or look-up movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a standardized recognition system is used, then the system structure is simple, but it cannot adapt to individual users with non-standardized speech and movement patterns
Solution Approach 1:
The recognition system is segmented into multiple independent classifiers, each specialized for a specific user or input type. This allows the system to handle individual user variations without requiring complete system redesign, resolving the contradiction between adaptability and complexity by distributing the adaptation burden across separate modular components.
Solution Approach 2:
The system performs preliminary classification to identify the user and input type before processing the actual recognition task. This preliminary action allows the system to select the appropriate trained classifier, enabling adaptation to individual users while maintaining a standardized overall system structure.
2Measurement precision
If a customized recognition engine with trained classifiers is implemented, then recognition accuracy for individual users improves, but the training time and system complexity increase
Solution Approach 1:
The system collects more training data samples than the minimum theoretically required, and processes all available samples through the training pipeline. This excessive action approach ensures that even users with limited or atypical speech/movement patterns achieve sufficient recognition accuracy, accepting longer training times as a trade-off for improved precision.
Solution Approach 2:
The system automatically performs the entire training process without requiring external intervention or manual adjustment. The classifier is self-trained using collected user samples, which reduces the time and expertise needed for system configuration while maintaining high recognition accuracy through personalized adaptation.
3Adaptability or versatility
If multiple classifiers for different input types are maintained, then the system can handle diverse speech and gesture patterns, but the device complexity increases
Solution Approach 1:
The system employs a universal classifier architecture that can process multiple input types (speech, facial gestures, head movements) through a common processing framework. This multi-functional design allows the system to support diverse input types while avoiding the complexity of completely separate processing pipelines for each input modality.
Solution Approach 2:
The system performs preliminary classification to identify the input type before routing to the appropriate processing path. This preliminary action organizes the complexity of handling multiple input types into a structured sequence, reducing the operational complexity while maintaining versatility.
4Reliability
If the system is highly customized for each user, then communication effectiveness improves, but the ease of operation and setup decreases
Solution Approach 1:
The system automatically performs user-specific customization through self-service training, eliminating the need for manual configuration by operators or therapists. Users simply provide sample inputs during a brief training period, and the system autonomously creates personalized classifiers, thereby maintaining high communication effectiveness while significantly improving setup ease.
Solution Approach 2:
The system performs all necessary customization and adaptation actions during an initial preliminary training phase. Once this preliminary action is complete, the system operates in a standardized manner without requiring ongoing manual adjustment, thus achieving both high reliability and ease of operation during normal use.
Data Source
AI summary
A customizable recognition system with at least one processor to process the audio/video input to determine a control command for accessibility functions of a computer or gaming application. The customized recognition engine has a classifier for each different input type for the different types of speech or gestures. The classifier stored with a link or indication of a user identifier. The interface is configured to provide the control commands to a computer application, gaming application, or a laptop, or an access technology device.