Semantic Segmentation Weighting for Minute Medical Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems using convolutional neural networks (CNNs) struggle to accurately identify minute objects such as the center point of a vertebral body, center points of right and left eyeballs, or the center line of an aorta in medical images.
Innovation Solution
A semantic segmentation model is trained using annotation images where specific pixels corresponding to objects are assigned a higher weight, allowing for accurate identification of one point, discrete points, or line structures in medical images by generating a feature amount map and applying non-maximum suppression processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a standard semantic segmentation model is used to identify objects in medical images, then large objects such as lung fields or spines can be identified with relatively high accuracy, but minute objects such as vertebral body center points, eyeball center points, or aorta center lines cannot be accurately identified
Solution Approach 1:
The patent applies local quality by assigning different weights to different pixels in the loss function. Specifically, pixels corresponding to minute objects (such as vertebral body center points, eyeball center points, and aorta center lines) are assigned higher weights than pixels corresponding to large objects. This allows the model to focus computational attention on critical minute structures during training, thereby improving identification accuracy without compromising overall reliability
Solution Approach 2:
The patent changes the parameter weighting in the loss function from uniform to non-uniform. By introducing a weight map where pixels corresponding to minute objects have higher weight values, the training process is modified to prioritize learning features of minute objects. This parameter change enables the model to achieve better measurement precision for small structures while maintaining reliable identification across all object types
Data Source
AI summary
A processor uses a semantic segmentation model that has been trained using an annotation image in which a first pixel corresponding to at least any one of one point corresponding to an object, a plurality of discrete points corresponding to a plurality of objects, or a line corresponding to an object having a line structure is set as a first pixel value and a second pixel other than the first pixel is set as a second pixel value different from the first pixel value, the model having been trained by assigning a greater weight to the first pixel than to the second pixel to calculate a loss, inputs an image to the model and outputs a feature amount map having a feature amount related to the one point, etc. in the image from the model, and identifies the one point, etc. in the image based on the map.