Image Resize Processing for Camera Distance Standardization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image classification accuracy deteriorates due to variations in image composition caused by different camera-to-subject distances, leading to inconsistent feature calculations for the same subject.
Innovation Solution
An image processing device that performs resizing of images to a predetermined camera-to-subject distance, calculates feature amounts, and makes category decisions based on these standardized features, using a classifier generated from sample images resized similarly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a classifier is used for image classification, then classification capability is provided, but classification accuracy deteriorates due to composition variations from different camera-to-subject distances
Solution Approach 1:
The patent applies parameter changes by resizing images to standardize their composition. Specifically, images are resized to match a predetermined camera-to-subject distance, transforming the spatial parameter of the image composition. This standardization allows the classifier to accurately distinguish between different categories without being confused by variations in camera distance, thereby resolving the contradiction between providing classification capability and maintaining high accuracy despite composition variations.
2Reliability
If images of the same subject are captured at different camera-to-subject distances, then diverse composition is achieved, but feature amount becomes inconsistent leading to reduced discrimination accuracy
Solution Approach 1:
The patent applies preliminary action by performing resize processing on images before they are used for classification. The system pre-processes images by resizing them to a standard composition corresponding to a predetermined camera-to-subject distance. This preliminary standardization ensures that when feature amounts are calculated, they are based on consistent compositions, thereby improving feature consistency and reliability without adding complex processing steps during the actual classification phase.
Data Source
AI summary
An image processing device includes: an input unit that inputs an image obtained by imaging; a resize processing unit that performs one of enlargement and reduction processing for the image input by the input unit as a category decision target image, so as to obtain a first image equivalent of a predetermined camera-to-subject distance; a feature amount calculation unit that calculates a first feature amount that indicates a feature of the first image resized by the resize processing unit; and a decision unit that makes a category decision on the category decision target image based upon the first feature amount, the category decision being to make a decision as to which of a plurality of categories the category target image corresponds to.