,
To those who have responded to my raising the issue of the current status and likely fit-for-purpose after the current clinical trials of the images produced by the EMV Gen #1 machine I say respectfully to all:
,
Stotes ….. your reply is that we have a machine learning model which should work, and the right people and organisations are applying themselves to making it work, and the people at EMV have made lots of commitments on the assumption that it will indeed work, but there has been no comment as to progress in regard to image quality. As you say “I believe that EMV should see improvements in their resolution/imaging. As to when, I unfortunately don’t have the answer”
,
Gareth: … I listened to the presentation by Mr Alexander and like you are very impressed but as you say EMV or our imaging project is not mentioned.
,
Vintage: … The article you provide says “a deep learning algorithm for microwave imaging is proposed” … the article is proposing that such an algorithm could/should be developed, which is exactly what EMV are trying to do.
,
A separate academic paper to the one you referenced, from the same group in Queensland, describes essentially what EMV’s Gen#1 unit looks like with lots of technical information and leads into the conclusion “are promising candidates … “
,
Diagnosisof a brain stroke using wide-band microwave scattering
Dastan Ismail
and
Samah Mustafa
Published:22 March 2023https://doi.org/10.1098/rsos.221560
Abstract
This paper presents a new computer-aided microwave monitoring system as a promising, portable and inexpensive tool to detect and localize brain stroke using a bank of a new wavelet-matched filters. The head is exposed to microwave radiation over the band from 1.1 to 3 GHz, and the backscattered signals at a hemi-elliptical array of 16 antenna elements surrounding the head are filtered to analyse the perturbation in the microwave signals from the brain. A novel technique is applied to remove the strong reflection from the air–skull interface as a way to estimate the target response and is compared with different techniques from literature to portray their role in the performance. The study results approve that the intensity and the distribution of wavelet energy and Shannon wavelet entropy in the filtered microwave signal, and the novel tool based on the distance between the wavelet energies at symmetric opposite antennas are promising candidate signaturesfor computer-aided detection and localization of a stroke.
,
A competitor in the field, the Hyperfine unit, has negative aspects that we know about that will make our unit much more practical to use, but also there is criticism of the usefulness of this machine’s images. From memory it doesn’t handle small (<2mm) bleeds.
,
I note also that EMV has recently hired additional people to the team (ref dysone post 14th April) who have particular technical skills relevant to development of EMV’s unit. Zili Xu, PhD, as Lead AI Engineer and Mostafa S as RF Systems Engineer. At this late stage EMV are hiring technical people who don’t come cheap and although it is of course great to have them on the team the fact that we need them could be a worry.
,
So in summary, great idea, great team, and obviously we are here because we see the benefits, but nowhere else in the world has it ever been done before.
,I would like to put more of my hard-earned into increasing my shareholding but the lack of any progress reports on the critical issue of image quality adds risk to my decision which is totally not necessary, and could be a possible explanation for the downtrend in SP.
.
Expand
(20min delay)