Since first introducing TBDx, to professionals in South Africa and India, we have referred to the technology as ’automated sputum microscopy’, or ASM technology. The reference was intended to quickly describe a fully automated diagnostic technology as applied to a normal laboratory procedure for the detection of tuberculosis, routine human visual interpretation of stained sputum slides as viewed under a microscope. The intended point has been, by adopting our technology, the community will be using familiar laboratory equipment that has been deeply embedded in their current operations for over 100 years. This ASM reference has been advantageous in helping laboratory healthcare professionals to quickly understand the TB diagnostic technology we offer today.
During our recent internal strategic discussions we have debated the use of this term and whether it accurately describes the genuine value of the TBDx technology. Certainly the microscope is an important vehicle to assist in the image acquisition process, as is the camera. However the well-focused images would not be quite as valuable without the computer algorithms that perform the detection analysis. Indeed, the automated focus is a critical element, and its development was perhaps the most important challenge that we successfully overcame during development.
Since its founding, our company has been involved in acquiring, analyzing, segmenting, extracting, and classifying objects of interest in images from multiple acquisition sources. This area of expertise is often referred to as computer vision or machine learning. The field focuses on the extraction and identification of pixel data that is extremely difficult or impossible for a human to perceive. In the company we use the terms computer vision and supervised machine learning interchangeably. In our view it is ’computer vision’ – the image processing and pattern recognition algorithms – that brings the diagnostic value to our automated detection solutions. The slide loader, barcode reader, microscope, automated slide stage and camera are all part of a finely integrated hardware platform, one that can be used for multiple tropical or infectious disease applications. But, it is the segmentation, the feature extraction, the pattern classification and the very sophisticated computer analytics and supervised software learning that makes the digital diagnosis possible.
We have decided to re-brand TBDx as ’computer vision diagnostics’. The intent is to communicate that our technology leverages software, digital pattern recognition and signal processing, pixel analysis, and other aspects of imaging science to arrive at a diagnosis.
As our technology evaluations move forward in South Africa and India, and you see more of this reference, you will better understand the reason for the change.