Name: Ubotica Technologies Ltd.
Date founded: 2016
Number of employees: 12
Location: Dublin, Ireland; Ciudad Real, Spain
Irish SME focused on delivering computer vision and AI based solutions on-the-edge.
Innovation Management Partner:
IDRD: In-Line Diabetic Retinopathy Detection
Diabetic Retinopathy (DR) is the leading cause of vision loss in adults aged between 25 and 60, with a recent study estimating that the number of people worldwide with DR will grow to 191 million by 2030. When DR is caught early, treatment is effective at reducing or preventing vision loss. Only a very small number of countries have national screening programmes for DR detection in place, and they are costly to run.
Ubotica has developed a deep-learning based solution for detecting the presence of DR indicators in retinal images on-the-edge, in the specialist fundus cameras that take these images. The solution has been designed to assist ophthalmologists to accurately and consistently evaluate and diagnose DR. To support this, it also provides information related to the basis on which the assessment was made.
A light weight solution has been developed around the embeddable Intel Movidius Myriad X Vision Processing Unit (VPU), which pre-processes the retinal images using its on-board image filtering capabilities and then applies an Artificial Intelligence Convolutional Neural Network (CNN) to assess the retinal image for DR indicators. Runtime analysis with the Neural Network Dependability Kit (NNDK) is applied to display to the user the CNN training images which most closely match the image being analysed. This introduces a level of confidence to In-Line Diabetic Retinopathy Detection that heretofore was not possible, adding a Verifiable AI element. The integration of the working prototype with a retinal imaging camera has been demonstrated.
A significant challenge in the deployment of AI based solutions is in the supervising of decisions of the CNNs which are at the core of these systems. Addressing this, fortiss supported in interpreting the NNDK metrics for assessing the performance of the CNN, also after a vigorous pruning process to reduce the model’s size and make inference faster, and in integrating the runtime analysis into the application. Depending on their specialization, all FED4SAE partners provided input to refine the technical approach and business case, and joint dissemination opportunities were exploited.
Ubotica are working with fundus camera manufacturers to introduce the solution to the marketplace, offering customization and design-in services for the VPU, CNN and associated software. The NNDK’s supervisory role on the decisions being made by the CNNs delivers a significant benefit to solution integrators. Deploying Ubotica’s solution in fundus cameras will significantly enhance the role of both the cameras and the ophthalmologists who use the cameras, with further advances anticipated with the introduction of detectors for other pathologies, such as Glaucoma and Hypertension, all based on the same underlying hardware.