07 - Swarm One: Vizyon demo by Nicolas Kritter & Pierre Durant

now we move on to to uh a demo from a project called vision and as they have footage swarm has helped help them to offer patients level of services they couldn't get otherwise vizion has been using swarm for quite quite quite some time now and has already been featured at the swarm alpha event so let's see what has happened since in a demonstration that we can now play please hello i'm nicola from vision we build a decentralized teleradiology platform using swab and gas to provide better access to medical expertise in france and worldwide i will first present you what we do with swarm and then show you a demo of our application first what is the teleradiology platform and what do we do our job is to connect health sector ensuring security integrity and confidentiality of the data we need health center ai providers and radiologists to get the best from the human and the machine in order to deliver a medical report to the patient our standard workflow is a patient goes to the center to get an exam after getting consent we will retrieve their data anonymize it and distribute it to specialized ai depending on the exam we call different ais then we provide the radiologist the result of the ai and the medical data to analytical data so he can write his own report finally we integrate the radiologists report in the health facility center system our challenges are that we have to integrate with very heterogeneous ecosystem with sometimes very legacy technologies machines are on different private networks and actors are geographically decentralized we started to build the platform in mid 2019 and launch it in production in end 2019 with a pre-beta swarm version we are using swarm and gas our core stack for now three years and participate participated at each launch event here is some figure on how much we use the platform so we have traded 17 thousand plus patients since the swamp since the last swarm beta launch event in february we plug in 12 different fat facilities and uh two hospitals that use our system for fast coverage free using ai-only workflow we integrated four different types of lcs and earth as system for automating processes finally a quick example of why using distributed system like swarm is essential when managing important data in march 2021 one of the data centers of ovh caused fire causing millions of whatsapp of website to go down losing data and back up forever many people thought it would be impossible but it happened we lost we lost half of our machines in overage in this event but not only our service has been merely impacted we ever lost any data no medicated no medical data were lost during this event now let's start the demo here is our main application we have a radiologist on the left side and the l center on the right side everything is recorded simultaneously and the applications are on two different machines to mimic the production environment first cl center creates a patient and then will add all the necessary medical data we try to automatize this step as much as possible by being plugged on their system usually the patient is already created with its image and the user just adds have to add the missing details the main block time button neck is because of the blockchain as we have a two second block timer so we have to find tricks to ideate from the user experience like unlocking the next action batching and so on then the earth center will send the data to the radiologist the patient appear now on the audiologist's work list he then can go on by downloading the images of the patient reading the prescription the the result of the ai and analyzing the images then you will write it report the radiology still have to download images as we need to process them inside its own software to add many features by for example changing contrast zooming and handling all this medical data finally he will send the report to the center while the passion is being sent the radiologist can go on to the next patient to avoid the time waiting for the transaction to be included in the chain the report is then received by the health center and will be integrated in their system thanks for watching so thank you nikolai npr for this inspiring presentation i think that when real life cases like these are presented it really shows the power and potential of the technology um so maybe you could uh just tune in for um a minute hi hello hello hi hello everyone hi hello so um nice to see you so what are your plans for the future and when when can people where can people get to know more about your project so about a project we have uh so you see our website uh vision.ai we have some linkedin but for another project is to continue to use more swarm using the the latest version and implementing more processes inside the inside the swarm environment for example the last as the last event is from beta we were talking about using satellite connection and integrity the dicom viewer and so we managed to use the satellite connection to connect the overall centers that are in france and sometimes a very very low bandwidth which is a very big problem for terror ideology because we are managing these files and so on the next project will be for example to have an integrated viewer so the artist will be able to see the images inside is one is on the iron margin work list and we are now working on some way to better implement the result of the ai to really uh create a free report for the radiologist so the arduous can really focus on just confirming what the ai said and add his own expertise as well so and so everything would be a insurance one to under all the logics and uh maybe we'll add uh handling more data than just uh as data but maybe all the data around to offer a better experience to the ideologist connect to outer world and other processes other actors and as well as center to manage their their environment because of swamp we have been able to deliver an incredible value to our patients and since the beginning we are close to many gigabytes maybe terabyte of medical that are shared with swarm and delivering value into care of patients right now we have some very interesting notes with very high latency or with starlink connected in the mountains and all of this use the same system the swarm system so we'll continue to use it and develop it and we have to connect with a huge ecosystem on the main network now that it's it's life so let's continue together and uh actually a few months ago we made an experiment with a s practitioner on the other side of the world in the institute in polynesia we have a node running there also in the data center also in hawaii to deliver um a good data throughput for the radiologists that are are doing some night hook uh in the other part of the world so uh its case it scales very well thank you that is very nice to hear um so uh we have to stress right that you have your own swarm sub network so you're not working currently on the main net uh but you have it uh um i think the the um it proves the technology what the technology is capable of nevertheless right so yes thank you thank you to all the team thank you very much thank you