Event-Driven Image Processing for Pixel-Change Triggered Computing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in processing vast amounts of real-time image data efficiently, particularly in applications where significant changes are infrequent, leading to unnecessary data transmission and high power consumption, especially in high-speed imaging and smart city sensor applications.
Innovation Solution
An event-driven device with a neural network that processes data only when pixel changes occur, utilizing analog multipliers and a comparator to trigger processing, allowing for prioritization of pixel changes and selective storage of frames for non-linear comparison, thereby reducing unnecessary data processing and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If real-time image data is processed continuously at high speed, then processing timeliness is improved, but power consumption and data transmission volume increase significantly
Solution Approach 1:
The system performs periodic comparisons between current frame pixels and reference pixels, but only triggers full processing when changes exceed a threshold. This periodic checking with conditional activation reduces power consumption while maintaining real-time responsiveness to significant changes.
Solution Approach 2:
The invention extracts only the changed pixels from the full image frame and transmits/processing only this subset. By separating and processing only the relevant portion (changed regions) rather than the entire frame, power consumption and data transmission are significantly reduced while maintaining processing timeliness.
2Measurement precision
If high resolution sensors are used to detect small changes, then detection precision is improved, but data volume and processing complexity increase
Solution Approach 1:
The image frame is segmented into individual pixels that are independently compared against reference values. This segmentation allows the system to process only the minimal subset of changed pixels rather than the entire high-resolution frame, reducing processing complexity while maintaining the detection precision of the full-resolution sensor.
Solution Approach 2:
The invention extracts only the changed pixel values from the high-resolution data stream and transmits only this extracted subset for further processing. This extraction approach maintains the high detection precision of the original sensor while dramatically reducing the complexity of downstream processing and transmission.
3Loss of information
If all pixel data is transmitted for analysis, then data completeness is improved, but transmission bandwidth and processing load increase
Solution Approach 1:
The system extracts only the changed pixel data from the complete frame and transmits only this extracted subset. This ensures that all necessary information (the changes) is transmitted while minimizing the total data volume, directly resolving the contradiction between data completeness and transmission quantity.
Solution Approach 2:
Instead of transmitting the complete frame data, the system performs a partial transmission of only the changed pixels. This partial action is sufficient to maintain data completeness for the purpose of detecting changes, while significantly reducing the transmission volume and processing load.
Data Source
AI summary
An event driven device has a network collecting data. A device is coupled to the network for determining changes in the data collected, wherein the device signals the network to process the data collected when the device determines desired changes in the data collected. In a second embodiment a level shift adjusts the band diagram of a spill and fill circuit to allow processing only if a change in input value occurs. This is extended to teach a means by which the subset of an image or incoming audio data might be used to trigger an event. It could also be used for always on operation at lower power than alternative solutions.