Biometric Ocular Measurements With Deep-Learning Landmark Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for measuring landmark dimensions in the human eye, such as anterior chamber depth and lens thickness, are highly dependent on the surgeon's skill set and suffer from inaccuracies, low speed, and lack repeatability.
Innovation Solution
Employing a deep-learning convolutional neural network (CNN) to automatically detect landmark features in eye images, followed by post-hoc processing to refine these locations and calculate accurate interocular dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional manual box thresholding and hard-coding of dimensional data are used, then the measurement process is simple to implement, but the accuracy and repeatability deteriorate due to high dependence on surgeon's skill set
Solution Approach 1:
The patent replaces manual mechanical measurement processes with an automated deep learning system. The convolutional neural network automatically detects landmark features and calculates dimensions, eliminating the need for manual box thresholding and hard-coding. This substitution maintains ease of implementation while dramatically improving measurement accuracy and repeatability by removing human skill dependence.
Solution Approach 2:
The system performs self-service by automatically detecting landmark features and calculating dimensions without requiring manual intervention. The deep learning model processes images autonomously, generating predictions that are then refined through post-hoc processing. This self-service capability maintains implementation simplicity while achieving consistent, high-precision measurements independent of surgeon expertise.
2Device complexity
If traditional manual measurement methods are used, then the system complexity is low, but the speed and productivity deteriorate
Solution Approach 1:
The patent applies preliminary action by pre-training the convolutional neural network on extensive datasets of eye images with annotated landmarks. This pre-training enables the system to perform rapid, accurate measurements without requiring complex real-time computation during actual use. The model's weights and architecture are prepared in advance, allowing fast inference while maintaining high productivity.
Solution Approach 2:
The system replaces slow manual measurement processes with automated deep learning inference. Once the neural network is trained, it can process images and generate dimension estimates rapidly, eliminating the time-consuming manual annotation and measurement processes. This substitution maintains manageable system complexity while dramatically improving measurement speed and productivity.
3Ease of operation
If manual landmark marking and measurement are performed, then the process is easy to control, but the repeatability deteriorates due to variability in surgeon skill set
Solution Approach 1:
The patent replaces manual landmark marking with automated deep learning detection. The convolutional neural network consistently identifies landmark features based on learned patterns from training data, eliminating variability introduced by different surgeons' skills and experiences. The system maintains ease of operation through automated processing while achieving high repeatability and reliability in measurements.
Solution Approach 2:
The system incorporates feedback through post-hoc processing that refines the deep learning model's initial predictions. The refinement process uses classical image processing techniques to correct and adjust landmark detections, ensuring consistent and reliable measurements. This feedback mechanism maintains operational simplicity while significantly improving repeatability by continuously optimizing the measurement results.
Data Source
AI summary
A method for estimating biometric landmark dimensional measurements of a human eye includes, in a possible embodiment, receiving one or more images of the human eye via a host computer. In response to receiving the one or more images, the method includes generating a preliminary set of landmark point locations in the one or more images via the host computer using a deep-learning algorithm, and then refining the preliminary set of landmark point locations using a post-hoc processing routine of the host computer to thereby generate a final set of estimated landmark point locations. Additionally, the biometric landmark dimensional measurements are automatically generated via the host computer using the final set of estimated landmark point locations. A data set is then output that is inclusive of the set of estimated landmark point locations. A host computer that executes instructions from memory to perform the method.