Smartphone Retinal Imaging Real-Time Panorama Stitching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Capturing wide-field retinal images is challenging due to the need for precise alignment and stability, which is difficult to achieve with handheld systems like smartphones, and traditional methods require post-image acquisition processing to construct panoramas, leading to potential inadequacies in image quality detection after patient exams.
Innovation Solution
A system and method using a smartphone camera with associated optics to capture a digital video stream of retinal images, process them in real-time to identify points of interest, determine overlapping alignments, and combine images into a panorama, allowing for real-time construction of a wide-field retinal image with quality assessment and discarding of lower quality frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If handheld smartphone-based imaging systems are used, then portability and cost are improved, but alignment precision and image quality are worsened
Solution Approach 1:
The system performs preliminary actions by capturing multiple video frames and identifying points of interest (blood vessels, optic disc) before final image assembly. This preliminary processing allows the system to pre-determine alignment relationships and compensate for handheld instability, achieving good alignment precision without requiring a stationary tabletop setup.
Solution Approach 2:
The system implements feedback by continuously analyzing video frames to track the position of retinal features and dynamically adjusting the panorama construction algorithm. This real-time feedback mechanism allows the system to compensate for hand movements and maintain alignment precision throughout the examination, resolving the contradiction between handheld portability and alignment accuracy.
2Area of stationary object
If wider field of view is implemented, then diagnostic capability is improved, but alignment precision requirements increase
Solution Approach 1:
The system segments the wide field of view into multiple overlapping video frames captured sequentially. Each frame is processed independently to identify points of interest, and then these segmented frames are stitched together using feature matching. This segmentation approach allows the system to achieve a wide overall field of view while maintaining alignment precision within each individual frame, as the alignment burden is distributed across multiple smaller, more manageable segments.
3Loss of time
If post-image acquisition panorama construction is used, then processing time is reduced, but image quality assurance is worsened
Solution Approach 1:
The system performs preliminary quality assessment during the image acquisition phase by analyzing each video frame as it is captured. Points of interest are identified and quality metrics are calculated in real-time, allowing the system to pre-select the best frames for panorama construction. This preliminary action ensures image quality assurance is maintained while minimizing post-processing time, as poor quality frames are identified and discarded during acquisition rather than requiring extensive post-exam evaluation.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A method, system or apparatus for generating a wide field retinal image. System embodiments comprise a camera, which may be a smartphone camera and associated optics to capturing a digital video stream of individual retinal images and store the video stream of individual retinal images in a memory associated with the camera. The system or method then implements the steps of processing each retinal image and combining selected retinal images into a panorama in real time, during an ocular examination.