Continuous Volumetric Skin Lesion Views for Remote Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Remote patient assessments, particularly in dermatology, face challenges due to poor-quality images acquired by non-experts, leading to unreliable diagnostic outcomes and inaccurate lesion risk assessments.
Innovation Solution
A volumetric representation generation system that guides patients to capture images from multiple angles, processes these images using a trained model to generate a continuous digital volumetric representation, and provides this representation to medical professionals for accurate diagnosis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If patient-acquired images are used for remote dermatology assessment, then accessibility and convenience are improved, but image quality and diagnostic reliability deteriorate
Solution Approach 1:
The system creates a synthetic 3D volumetric representation (digital twin) of the skin lesion from multiple 2D patient-acquired images. This virtual copy allows physicians to view the lesion from any angle and lighting condition, effectively copying the essential diagnostic information while eliminating the limitations of single-view 2D photographs.
Solution Approach 2:
The system transforms 2D images into a 3D volumetric representation, adding the dimension of depth and spatial orientation. This enables physicians to rotate, zoom, and view the lesion from multiple angles simultaneously, providing comprehensive diagnostic information that cannot be obtained from single-angle 2D images.
2Ease of manufacture
If single-angle 2D photographs are used, then acquisition simplicity is maintained, but diagnostic accuracy and lesion assessment reliability deteriorate
Solution Approach 1:
The system divides the diagnostic task into two parts: (1) simple patient-side image capture using standard smartphones, and (2) complex 3D reconstruction and analysis performed by the AI system. This segmentation allows patients to provide data easily while the system handles the complex processing to achieve high diagnostic accuracy.
Solution Approach 2:
The AI-based volumetric representation system acts as an intermediary between the simple patient-acquired images and the complex diagnostic analysis. It processes the 2D images through machine learning models to generate a standardized 3D representation that can be reliably assessed by physicians, bridging the gap between easy acquisition and accurate diagnosis.
3Loss of information
If multiple 2D images from different angles are collected, then view completeness is improved, but acquisition complexity and patient burden increase
Solution Approach 1:
The system provides automated feedback to patients by reviewing the quality and completeness of submitted images. Based on this feedback, the system dynamically adjusts instructions and requests additional images only when necessary, optimizing the balance between view completeness and acquisition complexity.
Solution Approach 2:
The system changes the parameter of image representation from multiple separate 2D photographs to a single unified 3D volumetric model. This transformation consolidates information from multiple angles into one comprehensive representation that can be viewed interactively, reducing the practical burden on patients while maintaining complete view information.
Data Source
AI summary
A method for generating a continuous digital volumetric representation of a region of a patient, comprising: (i) receiving, by a volumetric representation generation system, an instruction or request to generate a digital volumetric representation of the region of the patient; (ii) providing one or more guidance instructions regarding obtaining a plurality of images of the region; (iii) receiving the plurality of images, wherein at least of each of two or more of the plurality of images are obtained of the region of the patient from a different angle; (iv) processing, using a trained model, the received plurality of images to generate a continuous digital volumetric representation of the region; and (v) providing the generated continuous digital volumetric representation of the region to a user.