Image Classification Apparatus Quality Assessment Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face inefficiencies in classifying and managing low-quality images, leading to storage waste due to the time-consuming process of manually reviewing and deleting poor-quality images, which often results in many low-quality images being backed up unintentionally.
Innovation Solution
An image classification method and apparatus that evaluates image quality using characteristics such as definition, exposure, sharpness, and noise, categorizing images into quality categories for efficient saving or deletion, utilizing techniques like Tenengrad functions, frequency spectrum analysis, and human detection to automate the process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually review and delete low-quality images one by one, then image quality control is achieved, but user operation time is excessively long
Solution Approach 1:
The system automatically evaluates image quality using definition degree calculation functions (such as Tenengrad, frequency spectrum, or entropy functions) and executes deletion operations without requiring manual user review. The device serves itself by autonomously identifying and removing low-quality images based on preset quality thresholds, thereby eliminating time-consuming manual operations while maintaining quality control.
Solution Approach 2:
The patent replaces the mechanical manual review process with an automated image processing system that calculates definition degrees using algorithms (gradient functions, frequency spectrum analysis, or entropy calculations). This substitution transforms the manual quality assessment into an automated computational process, dramatically reducing operation time while preserving quality judgment accuracy through mathematical evaluation methods.
2Reliability
If users backup all images without review, then no images are lost, but storage space is wasted due to low-quality images
Solution Approach 1:
The system performs preliminary quality assessment of images before backup operations. By calculating the definition degree of each image using evaluation functions and comparing it against preset thresholds, the system identifies low-quality images in advance and excludes them from backup. This preliminary filtering action prevents storage space waste while ensuring that only quality-meeting images are preserved, maintaining reliability for necessary images.
Solution Approach 2:
The patent introduces quality parameter-based filtering (definition degree thresholds) to control the backup process. By setting preset threshold values for image quality parameters and dynamically adjusting which images qualify for backup based on these parameters, the system optimizes storage utilization. Images below the threshold are automatically excluded, transforming the backup process from unconditional to parameter-driven selective preservation.
3Productivity
If automated image quality evaluation is implemented, then operation time is reduced, but calculation complexity increases
Solution Approach 1:
The system implements different definition degree calculating functions suited for specific evaluation needs: gradient functions (Tenengrad) for edge-based assessment, frequency spectrum functions for texture analysis, and entropy functions for overall image information evaluation. By selecting appropriate local evaluation methods based on image characteristics and requirements, the system achieves efficient automated processing without excessive complexity, balancing speed and accuracy for different scenarios.
Data Source
Figure 1~2
Figure 3~4
Figure 5A~5B
AI summary
The present invention provides an image classification method, comprising: an analyzing step of analyzing the quality of any image; and a determining step of determining a quality category to which the any image belongs according to the analyzing result, wherein the quality categories include a first quality category recommended to be saved and a second quality category recommended to be deleted. Correspondingly, the present invention further provides an image classification apparatus. Through the technical scheme of the present invention, images can be classified according to the quality of the images, so that the user experience is improved.