Quantum Vision System Image Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current artificial vision systems face challenges with high processing times and large data sizes due to the vast amount of data in images, particularly in color images, and struggle with identifying features of low contrast.
Innovation Solution
A quantum vision system using noisy intermediate-scale quantum (NISQ) hardware and quantum machine learning techniques, combined with contrast enhancement and dimensionality reduction, to classify images efficiently, employing methods like contrast stretching, histogram equalization, and principal component analysis to reduce data dimensions and enhance feature visibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep convolutional neural networks are used for image classification, then classification capability is improved, but processing time increases significantly
Solution Approach 1:
The patent extracts and removes redundant information from images through dimensionality reduction techniques. By identifying and eliminating unnecessary features and data dimensions while preserving essential classification information, the system reduces the computational burden on neural networks, thereby decreasing processing time while maintaining classification accuracy.
Solution Approach 2:
The patent applies contrast enhancement selectively to specific regions or features within images that are most relevant for classification. By improving the quality and visibility of locally important features rather than uniformly processing entire images, the system enhances classification capability while minimizing overall processing time and computational resources.
2Measurement precision
If full-resolution color images are processed, then feature identification accuracy is improved, but data size and storage requirements increase
Solution Approach 1:
The patent extracts essential visual information from full-resolution color images by identifying and retaining only the most discriminative features. Through dimensionality reduction and feature selection, the system removes redundant data while preserving the critical information needed for accurate feature identification, thereby reducing data size and storage requirements.
Solution Approach 2:
The patent enhances contrast selectively in regions containing important features rather than uniformly processing entire images. This localized enhancement improves feature identification accuracy for critical areas while minimizing the processing and storage requirements for less important regions, effectively balancing accuracy with data efficiency.
3Productivity
If dimensionality reduction is applied to images, then processing speed is improved, but identification of low-contrast features becomes more difficult
Solution Approach 1:
The patent applies contrast enhancement as a preliminary step before dimensionality reduction. By pre-enhancing the contrast of low-contrast features in the original high-resolution images, the system ensures that these features are sufficiently prominent before any data reduction occurs. This preliminary action preserves the visibility of subtle features even after dimensionality reduction, maintaining identification accuracy while achieving processing speed improvements.
4Measurement precision
If contrast enhancement is applied to the entire image, then feature visibility is improved, but processing time increases
Solution Approach 1:
The patent applies contrast enhancement selectively to specific regions or features within images rather than uniformly processing entire images. By identifying areas containing important features and enhancing contrast only in those localized regions, the system improves feature visibility where needed while minimizing the overall processing time and computational resources required for images with large non-critical areas.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method of classifying an image using a quantum trained vision system is disclosed. The method comprises enhancing contrast of the image and applying dimension reduction to the contrast-enhance image. The enhanced contrast and dimensionally reduced image is passed to a quantum trained vision system and a result is generated from the quantum trained vision system. The result could be a simply yes/no or true/false binary result to see whether the image fell into one of several predefined classes. Alternatively, the result could be an indication of an object in the image.