On-Chip Neural Network Accelerator for Pre-Boot Accessibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user accessibility features in electronic devices are limited, particularly for visually impaired users, as they are often tied to the operating system and not compatible with all applications, and are unavailable in pre-boot modes like BIOS, with peripheral Braille displays being costly and limited to text conversion only.
Innovation Solution
A system-on-chip implementation that includes a touch controller, display controller, and neural network accelerator to provide audio and haptic feedback for graphical user interfaces, allowing for low-power, low-latency analysis of display content independent of the operating system, enabling users to interact with graphical elements and menus through touch events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If peripheral Braille displays are used for visually impaired users, then text accessibility is improved, but device cost and hardware complexity increase
Solution Approach 1:
The patent replaces mechanical Braille display hardware with a software-based solution that uses the existing touchscreen display and audio outputs to provide tactile and auditory feedback. The system captures touch event coordinates and displays regional information through software processing, eliminating the need for separate Braille hardware while maintaining accessibility functionality.
Solution Approach 2:
The patent makes the existing touchscreen display and audio system serve multiple functions: visual display, tactile feedback through region-based touch detection, and auditory feedback through speech synthesis. This multi-functional approach eliminates the need for dedicated accessibility hardware while providing comprehensive accessibility features across different applications and system modes.
2Ease of operation
If operating system-level accessibility features are implemented, then accessibility is improved, but compatibility with all applications and system modes is reduced
Solution Approach 1:
The patent introduces a software intermediary layer that operates between the user and the graphical interface. This intermediary captures touch events, determines relevant regional information, and provides accessibility feedback independently of the underlying operating system or application, ensuring consistent accessibility across all applications and system modes including pre-boot BIOS.
Solution Approach 2:
The patent segments the display into multiple regions and processes touch events at the region level rather than requiring full-screen or application-level accessibility implementation. This segmentation allows accessibility features to be applied locally to specific portions of the interface, making them compatible with diverse applications and system modes while maintaining effectiveness.
3Ease of operation
If comprehensive accessibility features are provided for graphical interfaces, then usability for visually impaired users is improved, but power consumption increases
Solution Approach 1:
The patent implements periodic action by processing accessibility feedback only in response to touch events rather than continuously monitoring or updating the interface. The system waits for user interaction, processes the relevant regional information and generates feedback only when needed, significantly reducing power consumption compared to continuous accessibility feature operation while maintaining full usability when required.
Data Source
AI summary
Computing platforms and related methods for providing accessible user interfaces are disclosed. An example apparatus includes a display controller to a display controller to sample a region of a display frame associated with a location of a touch on a display screen of a user device and logic circuitry to, after a Basic Input Output System (BIOS) is operational in the user device and prior to loading of an operating system of the user device, identify content in the display region and cause at least one output device to generate an output representative of the content, the output including at least one of an audible output or a haptic output.