Search This Blog

HearX Group- Artificial Intelligence Classification


Image result for hearx group otoscopeHear X is a start-up in South Africa, it was established in the year of 2015. Several audiologist and engineers came together to create affordable access to hearing healthcare using digital small solutions. The idea was to provide available products that would assess early onset of digital hearing problems among the population. The flagship technology, hearing loss and ear screening technology, it works with calibrated HD headphones. There is more than one million case that was performed in 39 countries. This is a rapid growing technological device that aims at providing patients with hearing impairment the best available hearing healthcare. Currently, in the clinical state, there are seven solutions in health which is a centralized data management platform.

HearScreen is a smartphone test performed using calibrated HD Headphones: this is used for quick hearing tests to determine if there is a hearing loss and suggest if the patient needs immediate medical care.

HearTest can be used to determine the threshold of hearing loss.

PeekAcuity is a visual acuity test.

hearTest Occ Health is the newest product. It is a certified digital, smartphone, audiometer. It gets used by industries to perform hearing tests, a baseline test can be set and hearing can be compared over time. According to the speaker, “employees seem to love the product, because they can go for a quick 30 minutes hearing test on the premise and thereafter get back to work without having to stop the work for a day.   

HearOAE: this is an otoacoustic emission screening and diagnosing tool used for newborn hear screenings; it is used to test newborns hearing to allow the health practitioner to treat the hearing impairment from an early stage onwards.  

HearScope was released through an Indiegogo campaign in the year of 2017. Indiegogo is a crowdfunded campaign.

Artificial Intelligence (AI) in hearScope: AI is the newest feature that was released in hearScope on the 9th March 2020. It has been under development since the year of 2017. hearScope is a smartphone video otoscope that can be used to view the “Tympanic Membrane” and ear canal. It captures images and videos of the Tympanic Membrane and ear canal; audiologists are excited about this product because in developing countries, public clinics and hospitals have a lot of patients and this can assist in triaging patients and ensure that patients receive the correct pre-examination treatment. It also enables patient involvement during an examination, because it enables clinicians to show the patient the images and videos taken of the ear canal and tympanic membrane.

The AI image classification feature assists healthcare practitioners to diagnose ear conditions. The AI model was trained using thousands of images that was all captured by the hearScope hardware. The images were  diagnosed by specialists; the system can diagnose wax obstructions, normal ears, chronic perforations, and abnormalities, which indicates a pathology is present and the patient should consult a specialist. As more images get collected, the AI model can be trained to recognize more pathologies. The AI is working towards being able to diagnose 12 pathologies, such as foreign bodies, cholesteatoma, acute otitis media, otitis media with effusion, myringosclerosis, wet chronic perforations, dry chronic perforations, grommets, wax obstructions, retractions, normal ears and unspecified. It is also planning it add extra pathologies to the system.

General practitioners are between 60-70% confident of diagnosing ear conditions. The AI image classification model has an accuracy of 94% to classify ear conditions, and it can assist health practitioners to confidently diagnose ear conditions. Furthermore, the AI plays a significant role in developing countries due to its low number of specialists, to assist hospitals Otorhinolaryngology (ENT) and general practitioners to screen both children and adults, to be able to triage and refer then accordingly. It helps to determine the type of ear disease to treat, for instance, if a patient complains of pain in the ear, the AI image classification can classify a wax obstruction and the patient can simply go to the clinic and a clinician can remove the wax blockage. The hearScope, with the AI image classification feature can be used by anybody and not only healthcare practitioners. 

A quick notice, with the current situation of Coronavirus (also known as COVID-19), there have been a high increase in inquiries of the AI image classification. Telemedicine has proven to be useful in a case like this; people can purchase a hearScope via the HearX website (www.hearxgroup.com), use the hearScope to self-screen and send the images and videos to their respective general practitioner (family doctor) for a diagnosis.  

The beta-version AI geature has been launched in March, the feature is currently free, but will become paid once the necessary FDA and CE approvals have been received. The AI classification works together with the hearScope hardware; the hardware has a cost incurred on it; the software is free. The hearScope hardware is 200 USD. The price for the AI subscriptions is not yet known.
The device has been under development since the year of 2017 when it was released on IndieGoGo. Various bodies contributed in the development of the product such as the advisory panel (four well known specialist across the world), research partners that assisted with the diagnosis of images, the image collectors, which captures the ear images. The images are validated and there is an exclusion criterion to the images, such as an image of the finger or any image that is not related to the ear. The validated images are further presented to the Panel of Research Diagnosis Partners (PRDP); the PRDP diagnose the situation based on their respective specialties. The hearScope hardware is FDA approved and meets all the hospital standard requirements.

The hardware is expected to be durable for at least 15 years depending on the number of usages on the hardware. The hardware id accompanied with a one-year warranty.
The biggest issue while developing the device, was integrating and improving the quality of the “Collection of Images”. Another problem that arose was finding specialists that can diagnose the images.

The application is also compatible with desktop devices such as Windows, Mac and Linux; however, the device smartphone app is only compliable with android phones. The hardware comes with a storage case for safe keeping. The hardware also has its own packaging such as a stant for the phone, adapter, specula which attaches to the front of the scope, and connectors.
The device is available to use and familiarize with. There are various institutions, such as the Manchester University in the UK or University of Pretoria in South Africa, across the world that have access to the device for academic purposes.

No comments:

Post a Comment

Thank you for your comments