Binary Scene Descriptor Generation With Low-Memory Convolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems struggle to efficiently generate descriptors from image sensors that highlight information of interest, particularly in applications requiring detection of predetermined elements and classification operations, without requiring extensive data storage.
Innovation Solution
An image sensor system that generates successive binary image planes and processes them on-the-fly to extract binary descriptors using convolution kernels, statistical functions, and logical functions, reducing memory requirements and enabling efficient detection and classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image sensors generate successive binary image planes to highlight information of interest, then detection and classification accuracy is improved, but data storage requirements increase
Solution Approach 1:
The patent extracts only the essential binary information from successive image planes by applying convolution kernels that detect specific patterns (edges, corners, textures). This extraction process converts multiple binary image planes into a compact set of binary descriptors that retain detection accuracy while dramatically reducing storage requirements.
Solution Approach 2:
The patent segments the image processing task into distinct stages: generating binary image planes, applying convolution kernels to extract features, and compiling binary descriptors. This segmentation allows the system to process and store only the essential extracted features rather than retaining all raw binary image data.
2Loss of information
If multiple binary image planes are processed to generate descriptors, then information of interest is highlighted, but processing time increases
Solution Approach 1:
The patent performs preliminary processing by generating binary image planes with specific bit depths and applying convolution kernels in advance. This preliminary action prepares the data in an optimized format that facilitates faster subsequent descriptor generation and reduces real-time processing time.
Solution Approach 2:
The patent employs periodic convolution operations with predetermined kernels applied to successive binary image planes. This periodic processing approach systematically extracts features at regular intervals, maintaining information retention while establishing a predictable processing rhythm that optimizes throughput.
3Measurement precision
If convolution kernels are applied to binary image planes, then feature extraction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent changes the parameter of the image data from multi-bit grayscale values to binary values (0 or 1) before applying convolution kernels. This parameter change simplifies the computational operations during convolution, as binary arithmetic is less complex than multi-bit arithmetic, while still preserving sufficient information for accurate feature extraction.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The present description relates to a scene descriptor generation system, comprising: - an image sensor (100) having a plurality of pixels (101), successively providing S binary image planes each having a unique binary value per pixel; and - a processing circuit (200) configured to, for each binary image plane provided by the sensor (100), implement the following successive steps: a) calculate N convolutions of the binary image plane by respectively N distinct convolution kernels, so as to provide N convolved images; b) generate at least one meta-image from the N convolved images provided in step a); and c) for each of the N convolved images provided in step a), generate a binary descriptor from said convolved image and said at least one meta-image.