Image Processing System Object Identification Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing systems struggle to accurately identify and distinguish specific objects in two-dimensional images, particularly when objects have similar shapes or appearances, such as doors and windows, due to limited pre-learned types and difficulty in distinguishing artistic shapes.
Innovation Solution
An image processing system that employs a processor to segment object blocks in a two-dimensional image, generate mask images, and input them into both a first and second image processing model to produce feature vectors and feature maps, which are then used to create weighted feature maps, enhancing object feature extraction and accuracy in object identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a known segmentation model is used to segment objects in a two-dimensional image, then the segmentation process can be completed, but the system cannot accurately identify specific object types or distinguish objects with similar shapes
Solution Approach 1:
The patent segments the image processing task into multiple components: first segmenting the image into object blocks, then generating mask images for each block, and finally processing these blocks through multiple image processing models to extract features. This multi-stage segmentation approach allows the system to handle diverse object types while maintaining identification accuracy for each specific object category.
2Measurement precision
If multiple image processing models are used to extract features, then object identification accuracy is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary feature extraction by generating mask images and extracting feature vectors from object blocks before final identification. The system pre-processes image blocks into standardized representations with extracted features, which can be efficiently matched against stored templates. This preliminary action reduces the computational burden during final identification while maintaining high accuracy.
3Measurement precision
If feature vectors are recalculated for each similar image, then processing accuracy is maintained, but processing efficiency decreases
Solution Approach 1:
The patent creates standardized feature representations (feature vectors) from image blocks that can be stored and reused. When processing similar images, the system can copy or retrieve previously calculated feature vectors instead of recalculating them from scratch. This copying mechanism maintains identification accuracy while significantly improving processing efficiency for similar or repeated images.
Data Source
AI summary
An image processing method includes the following steps: receiving a two-dimensional image and segmenting an object block in the two-dimensional image, masking the object block with a color block to generate a mask image; inputting the two-dimensional image and the mask image into a first image processing model, outputting a feature vector from the first image processing model; and inputting the two-dimensional image into a second image processing model to obtain a feature map. The feature map comprises a plurality of feature channel maps, and the feature vector contains a plurality of feature values. Each of the feature channel maps corresponds to one of the feature values in sequence. A weighted feature map is generated according to the feature channel maps and the feature values.