CMOS Image Sensor With Pixel-Specific Filters for AI Spectral Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CMOS imaging sensors sacrifice light sensitivity for color imagery due to the use of red, green, and blue color filters, leading to reduced light intake and increased blurring, which affects performance in low-light conditions and limits the quality of input data for artificial intelligence applications.
Innovation Solution
Implementing a CMOS imaging sensor with unique spectral filters for each pixel, configured to filter light based on a transmission function that differs for each pixel, allowing for higher spectral diversity and improved AI-efficacy by enhancing light sensitivity and spectral differentiation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If red, green, and blue color filters are applied to imaging sensor pixels, then color imagery is achieved, but light sensitivity is reduced
Solution Approach 1:
The patent divides the spectral filtering function into multiple segments by using multiple discrete spectral filters (e.g., red, green, blue, cyan, magenta, yellow) instead of a single continuous color filter array. This segmentation allows each pixel to receive light through different spectral filters, enabling both color information capture and improved light sensitivity through the multi-spectral approach.
Solution Approach 2:
The patent transitions from traditional spatial color filtering (using red-green-blue filters) to a multi-dimensional spectral filtering approach by incorporating additional spectral dimensions (cyan, magenta, yellow filters). This dimensional expansion in the spectral domain enables richer color information capture while maintaining higher light sensitivity through the increased number of spectral sampling points.
2Adaptability or versatility
If red, green, and blue color filters are applied to imaging sensor pixels, then color photography is enabled, but blurring of fine details increases
Solution Approach 1:
The patent applies different spectral filter characteristics to different regions or types of pixels within the sensor array. By assigning specific spectral filters (e.g., red, green, blue, cyan, magenta, yellow) to different pixel locations, the system achieves localized spectral optimization that preserves fine details while enabling color photography, as each pixel region is optimized for its specific spectral response characteristics.
3Adaptability or versatility
If conventional color filter arrays are used, then digital color imaging is achieved, but spectral diversity for AI applications is insufficient
Solution Approach 1:
The patent designs a multi-spectral filter array that serves multiple functions simultaneously: it enables digital color imaging through the red-green-blue filters while also providing enhanced spectral diversity for AI applications through the additional cyan, magenta, and yellow filters. This universal filter array design allows the same hardware to fulfill both conventional photography and advanced AI/ machine learning requirements.
Solution Approach 2:
The patent extends the spectral sampling dimension by adding multiple discrete spectral filters beyond the traditional three (red, green, blue). By incorporating five additional spectral filters (cyan, magenta, yellow and their variations), the system creates a high-dimensional spectral feature space that provides rich spectral diversity for AI applications while maintaining full color imaging capability.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution results in an imaging sensor that maintains high light sensitivity while providing richer spectral diversity, improving the quality of input data for AI applications and enhancing the accuracy of AI outcomes.
Implementation Method 1
Individual pixels of an imaging sensor may include a light sensor that produces an electronic signal based on the number of photons, or light intensity, that it receives
Implementation Method 2
spectral filters, each having a different transmission function, are placed in front of the pixels
Data Source
AI summary
An imaging device capable of producing images or data with relatively high spectral diversity, allowing for creation of information-rich feature vectors, is provided. Among other things, such information-rich feature vectors may be applied to a range of artificial intelligence and machine learning applications. The imaging device may include a substrate having a baseline spectral responsivity function, multiple pixels forming a cell fabricated on the substrate, and spectral filters each configured to filter light based on a transmission function corresponding to a substantially broad portion of the baseline spectral responsivity function. The spectral filters may be notch filters. Each of the multiple pixels in the cell may be configured to receive light through each of the spectral filters. The transmission function of each of the spectral filters may be substantially different for each of at least a majority of the multiple pixels in the cell.