Pixel Matrix Search Circuit for Low Power Outlier Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting a luminous element in a matrix of pixels are energy-intensive and complex, particularly when the element moves quickly, as they require high bandwidth and energy consumption to process large data streams, and are inefficient in mobile devices where power management is crucial.
Innovation Solution
A search circuit and method that divides a data matrix into lines, performing a first test to determine if any line contains data in a first state, sending compressed codes to indicate the presence or absence of such data, and a second test to identify specific nodes in the first state, reducing energy consumption and processing time by eliminating unnecessary tests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If recursive matrix division is used to search for outliers, then the position of outliers can be determined, but the process becomes extremely complex and energy-intensive
Solution Approach 1:
The patent divides the pixel matrix into multiple lines and processes each line independently through two-stage testing. This segmentation transforms the complex recursive matrix division into simpler, parallel line-based processing, reducing overall system complexity while maintaining outlier detection capability.
Solution Approach 2:
The patent extracts only the necessary information from each line by performing a first test to identify lines containing outliers, then performing a second test only on those specific lines. This extraction approach avoids processing the entire matrix with full complexity, reducing energy consumption and simplifying the search process.
2Speed
If high bandwidth search system is used to process large data stream, then quick position determination is achieved, but energy consumption becomes substantial
Solution Approach 1:
The patent performs a first test on all lines to identify which lines contain outliers, then performs a second detailed test only on those specific lines. This partial action approach processes only the necessary portion of data at full detail, reducing overall energy consumption while maintaining fast position determination for outliers.
Solution Approach 2:
The patent extracts and processes only the lines containing outliers with full bandwidth, while summarizing or skipping lines without outliers. This selective extraction reduces the effective data volume requiring high-bandwidth processing, thereby reducing energy consumption while maintaining speed for relevant detections.
3Use of energy by moving object
If data compression is applied to reduce data rate, then power consumption is limited, but processing complexity increases
Solution Approach 1:
The patent segments the data processing into two distinct stages: a first test that produces compressed line-level results, and a second test that processes only necessary pixel-level data. This segmentation allows compression at the line level while maintaining detailed processing only where needed, balancing power consumption and processing complexity.
4Measurement precision
If all lines are tested individually to find pixels in first state, then accurate position is determined, but processing time increases
Solution Approach 1:
The patent performs a preliminary first test on all lines to identify which lines contain pixels in the first state before performing detailed second tests. This preliminary action filters out lines without outliers, so the time-consuming second test is executed only on relevant lines, reducing total processing time while maintaining accurate position determination.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A pixel matrix is organized into rows of pixels (L). Each pixel is either in a first state or a second state. The matrix contains mostly pixels in the second state. Each row (L) of pixels is tested to determine whether or not it contains a pixel in the first state. The result of this test for each row is stored in a receiver (3). Rows (L) with at least one pixel in the first state are analyzed in more detail to determine the position of that pixel or those pixels within the row (L).