Dental Panoramic Image Artefact Correction via ML Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dental panoramic imaging systems face challenges in producing high-quality images due to incorrect patient positioning during x-ray data acquisition, leading to artefacts such as asymmetry, shadows, geometric distortions, and out-of-focus depiction.
Innovation Solution
A method and system that utilize a trained machine learning model to analyze image data from panoramic scans, generate a positioning parameter set indicative of the patient's head position, and reconstruct corrected-positioning dental panoramic images based on this data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If positioning and fixation means are used to reduce incorrect patient positioning and motion during acquisition, then image quality is improved, but patient comfort deteriorates and patients may react by positioning themselves incorrectly or moving during the scan
Solution Approach 1:
The system performs preliminary positioning assessment by capturing images during the panoramic scan and using a machine learning model to determine the actual patient positioning before reconstruction. This allows the system to pre-calculate correction parameters that compensate for positioning errors without requiring physical repositioning or fixation devices during the scan.
Solution Approach 2:
The invention replaces mechanical positioning and fixation means with a computational approach. Instead of using physical restraints to enforce correct positioning, the system uses machine learning-based image analysis and mathematical correction during reconstruction to achieve accurate images without mechanical intervention, thereby improving patient comfort while maintaining image quality.
2Measurement precision
If detailed guidelines and technician efforts are used to position patients correctly, then positioning accuracy is improved, but cases with head misalignment still occur due to patient movement or incorrect positioning
Solution Approach 1:
The system implements feedback by capturing images during the panoramic scan, analyzing them with a machine learning model to determine actual patient positioning, and using this information to adjust correction parameters during reconstruction. This closed-loop approach ensures that even if patients move or are initially positioned incorrectly, the system detects and compensates for these errors, maintaining consistent image quality.
Solution Approach 2:
The invention dynamically changes reconstruction parameters based on detected patient positioning. The machine learning model outputs positioning parameters that are used to adjust the reconstruction process, allowing the system to adapt to different positioning scenarios and maintain reliability across varying conditions.
3Manufacturing precision
If machine learning models and image analysis are used to determine patient positioning and reconstruct corrected images, then compensation for incorrect positioning is improved, but computational complexity and processing time increase
Solution Approach 1:
The system applies partial action by using a streamlined machine learning model that processes only the essential features needed for positioning assessment. Rather than performing exhaustive analysis of all image data, the model focuses on key positioning indicators, reducing computational complexity while maintaining sufficient accuracy for correction purposes.
Data Source
AI summary
A method for producing a dental panoramic image includes recording one or more images during a panoramic scan of a patient, providing an image data input to a trained machine learning model, the image data input being based on the recorded one or more images and the image data input being indicative of patient positioning during the panoramic scan, generating a positioning parameter set by the trained machine learning model, the positioning parameter set being based on the received image data input, where the positioning parameter set is representative of the position of the patient's head during the panoramic scan, reconstructing one or more corrected-positioning dental panoramic images based on x-ray frame data recorded during the panoramic scan and on the positioning parameter set. Further, an x-ray imaging system and a computer-implemented method.