Object Identifier Using Integral Image Pixel Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating integral images to identify objects in images are inefficient in reducing data volume while maintaining accuracy, with Japanese Patent No. 4582079 achieving only a 7% reduction in data by removing one bit from 15 bits.
Innovation Solution
An object identifying apparatus that generates integral images by successively integrating pixel values and acquires compressed data by extracting every other pixel, allowing for reduced data volume without losing original information, and uses this compressed data to identify objects with high accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If integral images are generated using all pixels from the original image, then object identification accuracy is maintained, but data volume becomes excessively large
Solution Approach 1:
The patent extracts only specific pixels (e.g., every other pixel or pixels at regular intervals) from the integral image to create a downsampled version. This extraction process reduces data volume by retaining a subset of pixels that still contain sufficient information for accurate object identification, thereby resolving the contradiction between maintaining accuracy and reducing data size.
Solution Approach 2:
The patent performs preliminary downsampling of the integral image before using it for object identification. By pre-processing the integral image to reduce its resolution while preserving key features, the system prepares compressed data that maintains identification accuracy but occupies less storage space and requires less processing power.
2Quantity of substance
If data compression is applied to integral images, then data volume is reduced, but object identification accuracy deteriorates
Solution Approach 1:
The patent applies different processing strategies to different regions of the integral image. By identifying and preserving pixels in critical regions (such as regions containing object boundaries or high-contrast features) while allowing more aggressive compression in less critical areas, the system maintains identification accuracy in important regions while achieving overall data reduction.
3Measurement precision
If all pixels are processed for object identification, then identification accuracy is maximized, but processing time increases
Solution Approach 1:
The patent extracts a subset of pixels from the integral image at regular intervals (e.g., every second or third pixel) to create a downsampled representation. This reduction in pixel count directly decreases the computational load for subsequent object identification operations, thereby reducing processing time while maintaining sufficient accuracy for practical applications.
Solution Approach 2:
The patent processes only a partial set of pixels (a downsampled subset) rather than all pixels in the original integral image. This partial processing approach achieves sufficient identification accuracy for most practical purposes while dramatically reducing the computational time and resources required, representing a deliberate trade-off between complete processing and efficient partial processing.
Data Source
AI summary
An object identifier is provided with a data-forming unit and a data-retrieving unit. The data-forming unit forms an integral image having an image region of the same size as an original image and an image to be integrated. The data-retrieving unit obtains compressed data derived by extracting, from a plurality of pixels constituting the integral image, at least every other pixel in at least one direction.