GoCheck Kids Leverages Machine Learning to Enhance Pediatric Photoscreening

GoCheck Kids (GCK) is a clinically validated, user-friendly, comprehensive photoscreening and visual acuity app that helps prevent vision impairments, blindness, and vision-related learning challenges in children aged 1-18. The app enables healthcare practitioners to screen for amblyopia (aka 'lazy eye') risk factors and other vision risks by providing an integrated, affordable solution in a simple smartphone app. The company currently serves over 6,500 pediatric teams in the US and Europe. The app is FDA-registered and CE certified. GCK developed a next-generation clinically validated, pediatric photoscreening platform in 2013. The platform consists of a smartphone app to perform vision screening and a backend hosted in the cloud. It enables healthcare practitioners to screen for amblyopia and other vision disorders in children in a fast, easy, and affordable way.

GoCheck Kids (GCK), a comprehensive photoscreening and visual acuity app, was seeking to enhance the image classification component of its pediatric photoscreening application through machine learning. The cloud-based system was data-rich, and it was hypothesized that machine learning could improve the app's usability by supplementing image analysis and improving user actions to capture the best image possible. GCK needed a robust and resilient machine learning infrastructure to run experiments on a dataset of over one million images faster and more cost-efficiently, to train a highly accurate image classification model for the application. Additionally, GCK wanted to alert the user in real-time to retake the image when the image was captured while the child was not looking directly at the camera. This required a machine learning infrastructure to prepare data, run experiments on the dataset, and build and train new image analysis and classification algorithms faster and more efficiently.

In collaboration with Amazon Web Services (AWS), GCK identified Provectus as a partner with the necessary domain experience to execute the experiment. Provectus reviewed GCK's image classification software, prior machine learning methods, and pipelines. They delivered a secure, auditable, and reproducible machine learning training infrastructure deployed on AWS. This included a robust experimentation environment with experiment tracking and model versioning, and active learning pipelines for continuous data re-labeling and re-training. Provectus also integrated the machine learning solution into the mobile application and into GCK's business workflow. The machine learning infrastructure development and implementation aligned with GCK's broader strategy and future product developments. It incorporated a feature store, experiment tracking, and machine learning pipelines. Machine learning model re-evaluation pipelines, as well as active learning pipelines for continuous, efficient data re-labeling and re-training were implemented to support ongoing model retraining, evaluation, tuning, and improvements as new labeled data and user feedback arrived.

• The new machine learning infrastructure on AWS supported regular model retraining, evaluation, tuning, and improvements to machine learning models as new labeled data and feedback arrived. More efficient data preparation and faster experimentation enabled GCK to increase the recall of machine learning models by 3X while preserving its precision. This brought in both short and long-term benefits, including improved usability of the app and increased customer satisfaction, by decreasing the final results of 'child not looking' by asking the user to retake the image. The robust and resilient infrastructure for machine learning bolsters GoCheck's quick, sustainable growth in the vision screening market. The GCK application is now equipped with advanced machine learning algorithms that further enable accurate and convenient photoscreening for a wide range of conditions, including amblyopia. Better access to affordable eye screening with GCK means that millions of children will be diagnosed in time, and will not suffer from vision impairments and blindness as adults.

• Increased the recall of machine learning models by 3X while preserving precision

• Conducted over 100 large-scale experiments in three weeks by three machine learning engineers

• 95% of machine learning engineers' time is now exclusively dedicated to experimentation