Image Block Classification Using Color and Frequency Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition technologies face challenges in accurately identifying landscape subjects without specific shapes and require hardware like sensors and GPS, making them difficult to implement in compact digital cameras or camera phones, and are inefficient due to large template data requirements.
Innovation Solution
An image identification method that classifies image blocks using color space information and frequency components, learning separating hyperplanes from training data to determine categories without hardware configuration, and utilizes consistency checking to validate category classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If template matching or feature point extraction is used for subject recognition, then recognition capability is provided, but accuracy for landscape subjects without specific shapes deteriorates and template data becomes too large
Solution Approach 1:
The patent changes the parameters for subject recognition by using color space information (hue, saturation, brightness) and frequency components instead of traditional template matching. This allows the system to recognize landscape subjects based on their visual characteristics rather than requiring exact template matches, thereby improving accuracy for subjects without specific shapes while reducing template data requirements
Solution Approach 2:
The patent replaces the mechanical template matching system with a software-based feature extraction and classification system. By substituting the need for physical template storage and matching with computational analysis of color and frequency features, the system achieves better performance for landscape subjects without the limitations of template-based approaches
2Measurement precision
If hardware such as sensors and GPS is used to estimate camera mode, then estimation accuracy is improved, but device complexity and cost increase making embedding difficult into compact cameras
Solution Approach 1:
The patent extracts the camera mode estimation function from hardware-dependent systems and implements it through software image processing alone. By taking out the requirement for sensors and GPS and replacing it with image-based feature analysis, the system maintains estimation capability while dramatically reducing device complexity and enabling embedding in compact cameras
Solution Approach 2:
The patent creates a software-based copy of the camera mode estimation function that replicates the capabilities of hardware systems without requiring the actual hardware. Through image processing algorithms that analyze color and frequency features, the system copies the estimation functionality achieved by sensors and GPS, thereby avoiding the complexity and cost of physical hardware
Data Source
AI summary
An image identification method for classifying block images of input image data into one of predetermined categories; the method includes the steps of: dividing image data into multiple blocks to produce block images, processing the feature quantity of each block image by their color space information and frequency component, learning separating hyperplanes that indicate boundaries of each category by reading in training data image that have labeled categories for each block and processing image feature quantity for each block of an training data image, and classifying respective block image to a category according to the distance from the separating hyperplane of each category for a newly acquired image to obtain the image feature quantity of block images. An imaging apparatus implementing the image identification method noted above is also disclosed.