Image Signal Processing with Nonlinear Pixel Code Downsizing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sensor devices generate large amounts of data, leading to inefficient computation and increased power consumption when used in machine vision systems.
Innovation Solution
An image signal processing device that downsizes image data by generating non-linear pixel codes through non-linear mapping and masking operations, reducing the code length to improve computation efficiency and reduce power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing image sensor devices generate and process a larger amount of data, then image data quality and human eye recognition capability are improved, but computation load increases and processing speed decreases
Solution Approach 1:
The patent segments the pixel code into multiple bit positions and selectively masks certain bits based on illuminance conditions. This segmentation allows the system to process only the necessary bit information for machine vision tasks, reducing computation load while maintaining sufficient data quality for object identification, recognition, detection, and motion detection functions.
Solution Approach 2:
The patent applies different processing quality to different bit positions based on local illuminance conditions. In low-illuminance environments, more bit positions are masked to reduce data volume, while in high-illuminance environments, fewer bits are masked to preserve quality. This local quality adjustment optimizes the balance between data quality and processing efficiency for machine vision applications.
2Measurement precision
If existing image sensor devices generate and process a larger amount of data, then image data quality is improved, but power consumption increases
Solution Approach 1:
The patent segments the pixel code into multiple bit positions and selectively masks certain bits based on illuminance conditions. This segmentation allows the system to process only the necessary bit information for machine vision tasks, reducing computation load and consequently reducing power consumption while maintaining sufficient data quality for object identification, recognition, detection, and motion detection functions.
Solution Approach 2:
The patent applies partial processing by masking only the necessary bit positions rather than processing all bits. This partial action approach reduces the computational workload and power consumption while maintaining adequate image data quality for machine vision applications, avoiding the excessive processing that would occur if all data were processed regardless of actual needs.
3Productivity
If non-linear mapping is performed on pixel codes, then computation efficiency is improved for machine vision, but code length increases
Solution Approach 1:
The patent segments the pixel code into multiple bit positions and applies non-linear mapping selectively to relevant bit positions. This segmentation allows the system to improve computation efficiency for machine vision tasks while controlling the overall code length by masking unnecessary bit positions after the non-linear mapping operation.
Solution Approach 2:
The patent changes the parameter representation from linear to non-linear through mapping operations on specific bit positions. This parameter change enhances computation efficiency for machine vision algorithms while the subsequent masking operation controls the final code length, achieving an optimal balance between processing efficiency and data size.
Data Source
AI summary
An operation method of an image signal processing device includes receiving a raw pixel code from an external image sensor device including an image pixel, generating a non-linear pixel code by performing non-linear mapping on the raw pixel code, and outputting a downsized pixel code obtained by performing a masking operation on bit positions including a first bit position corresponding to a most significant bit of the non-linear pixel code and a second bit position corresponding to a least significant bit of the non-linear pixel code. An amplitude of the raw pixel code is linear to illuminance of a light incident onto the image pixel, an amplitude of the non-linear pixel code is non-linear to the amplitude of the raw pixel code, and the non-linear pixel code has a first code length, and the downsized pixel code has a second code length smaller than the first code length.