Image Processing Registration for Accurate Pancreatic Lesion Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition and classification methods fail to effectively generate pseudo images of lesions, particularly for early-stage pancreatic cancer, due to the lack of consideration for the positional relationship between images and the inability to combine images accurately, leading to insufficient data patterns and challenges in early detection.
Innovation Solution
An image processing apparatus that acquires and registers images with registration information to generate synthetic images by combining a current image with a prior image, focusing on the pancreas region to create pseudo images of early-stage pancreatic cancer, using registration information to align and combine images based on region information and combining ratios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If images are simply superimposed without considering positional relationships, then the process is simple, but the pseudo image cannot accurately represent the target organ and combines regions outside the target organ
Solution Approach 1:
The patent performs registration processing before image combination to establish the positional relationship between images. This preliminary action of aligning images based on anatomical landmarks and structural features ensures that subsequent combination operations work with correctly positioned images, preventing the inclusion of extraneous regions while maintaining processing feasibility
Solution Approach 2:
The patent introduces registration information as an intermediary element that mediates between the first and second images. This registration information, derived from anatomical landmarks and structural features, serves as a bridge to correctly align the images before combination, ensuring accurate pseudo image generation without directly complexifying the overall process
2Productivity
If images are combined without registration information, then the processing is faster, but the positional relationship is not considered leading to inaccurate pseudo images
Solution Approach 1:
The patent performs registration processing as a preliminary step before image combination. By pre-establishing the positional relationship through registration based on anatomical landmarks and structural features, the system ensures accurate pseudo image generation while maintaining efficient processing through optimized registration algorithms
3Device complexity
If pseudo images are generated without considering anatomical structures, then the generation process is simpler, but the pseudo images cannot accurately simulate early-stage pancreatic cancer
Solution Approach 1:
The patent applies local quality by focusing registration and combination operations on specific anatomical regions of interest, such as the pancreas. By concentrating computational resources on relevant anatomical structures rather than processing the entire image uniformly, the system achieves accurate pseudo image generation for early-stage pancreatic cancer simulation while controlling overall system complexity
Solution Approach 2:
The patent performs preliminary registration processing that identifies and establishes the positional relationship of anatomical structures before image combination. This preliminary action ensures that subsequent combination operations accurately represent the target organ's anatomy, enabling reliable pseudo image simulation of early-stage pancreatic cancer
Data Source
AI summary
An image processing apparatus according to the present disclosure includes: an image acquisition unit configured to acquire a first image which has an abnormal region in a first region, and a second image which is different from the first image and includes at least the first region, by imaging at least one subject; a region information acquisition unit configured to acquire region information on the abnormal region in the first image; an registration information acquisition unit configured to acquire registration information related to registration of the first image and the second image; and a synthetic image generation unit configured to combine the first image and the second image based on the region information and the registration information, and generate a synthetic image which has a region corresponding to the abnormal region in the first region.