 Thank you for coming. I'm glad to be here and Really want to thank the organizer Linux Foundation organizing the whole event Last night you gotta actually get a chance to go to Guinix storehouse No Yeah, yeah, it was great event right so I love the third and I Love the audience entertainment and also the drum performance. It's great and the interesting the most interesting part to me is actually the Meeting with all the people last night. I met with so many different company different people and Interestingly when I sit down for dinner And the lady was sitting next to me and I kind of talk to him talk to her She actually came from Canada Canada working for a university and she is a professor And she's working on a health care project kind of similar to what I'm doing And she is kind of solving a different aspect of the problem Which is our security protecting data from a smartwatch and then how to how is that to get sent to a server? so after the dinner discussion was very interesting we promised to keep in contact and see if there's a chance that We can work together after the conference so So Today I want to make this a little bit more interesting for the audience So I prepare some gift first more gift and I'll ask Three questions at the end if you can correctly answer them you can you can get one of the gift so pay attention So today my talk is about recover monitor. So basically edge-based open source wearable devices how you actually capture all the biomarkers from the patient and then basically send it over to the doctors and running certainly I engines able to analyze whether the patient is actually on a very good recovery path And then with precision and efficiency. So that's basically the talk of today the entire process open source So before we get into the technical details. I have another my and another of my colleagues poundu He unfortunately is not able to come on site, but he prepared video So I'll play that video and he'll talk about the architecture for the solutions and also give a demo So but before that, let me give you a quick overview of what I'm talking about. So the remote patient monitoring market trend This is a huge market and the American Heart Association basically defined this which is a set of Teddy health and facilitate patient monitoring as well as transfer patient generated health data and Then basically helping the healthcare team to basically manage these patients in a timely manner So in terms of user Very interesting stats basically double between 2020 to 2025 to about 70 point six million Which is huge and they're looking at the market in terms of revenue generated is projected to generate 175 Billion in 2027 so if you look at the diagram kind of bar chart on the bottom of that green bar Roughly about one fourth is about devices revenue and then about three fourth is actually software and services We kind of kind of make sense A lot of investments actually need to be made to build the software infrastructure Which is what kind of what we're trying to do here as well And then according to a spy glass consulting a group nine of ten healthcare providers already testing and Considering testing remote patient monitoring. It's actually 88 percent And why they are so interested because these these are all the three major benefits actually highlighted by Kathman's Number one improved patient outcomes Which of course is the concern of all the doctors out there how to improve the patient outcome of their patients And the second is improve compliance rates Basically making sure their methods they deliver the infrastructure the data everything is compliant to the industry standards And then the third one is also quite important is the patient actually want to know more about the health so all these Services basically giving them information the history the data and they give them a view of how their health is progressing over time And they will be able to take additional actions So so let's let's drill into the smart watch Market particularly because we'll be using smart watch in this research And I'm actually wearing myself right here, and I'll show you a little bit later. So According to counterpoint they're forecasting Very healthy growth in fact last quarter Q2 2022 Actually grow about 13 percent year over year And one market actually is highlight here, which is India growing about 300 percent over 300 percent Which is significant and Apple is leading as you can see 29 from 3% Samsung second maintain a very healthy growth 40% year over year growth and Huawei actually Number one spot in China for several quarters already already and then again India Which is a very very high growth for you have to 347% What is interesting is if you look at this data India market Actually in Q2 2022 about 30% of the smartwatch they're selling is actually below 50 dollars So a very low price point and that's why they're able to basically jump in numbers in sales and These are the two vendors actually the top ones in India right now The fire boats and also noise If you look at actually one quarter before in Q1, they are not even there So within a very short time they're able to jump up to number four and number five in the overall market And why why are we using a smartwatch? I think one of the main reason is very affordable as I mentioned earlier the India market The cheapest version that they are selling very well is actually below 50 dollars A lot of these stats are pretty much supported by Most of the smartwatch that you can find in the market today. We're talking about body temperature Blood oxygen heart rate stress level or sleep pattern. So these are quite common But what is interesting is if you look at some of the leading brands leading product out there in the market They are innovation actually being made by a lot of this company and I will actually want to highlight One product and one research that might be very interesting for you One is actually blood pressure So this company actually created a product that able to measure blood pressure of the patient of the person actually wearing it So this is one of the watch actually can do a blood pressure So if you look at this watch is actually a little bit big but if you look at the wristband, this is actually Air pump so the watch can pump air and put pressure on my wrist and it will be able to measure blood pressure And I actually test it on myself and then I use a real blood pressure Device and then compared to the watch actually pretty accurate. So so this is one of the very interesting way how People patients can actually use a normal regular consumer devices and able to get more insight into the health And they're looking at a very interesting research if you look at the bottom Animation pictures This is actually done by researchers in UC, San Diego in California So what they did is actually they came up with a device a very tiny device You can see this person wearing on the arm With the lots of hundreds of these micro needles these micro needles Very very thin about one-fifth of a human hair They're able to basically Detect the bio molecules right underneath the skin. So interstitial Fluid under the skin and able to basically detect three things good cause level alcohol level and lactate level And these are all interrelated. So that's a very important for doctors or the patient to actually keep track of these How they changes over time? If you look at this particular device, what was interesting in this device is is not Penetrable, I mean, there's not something that built into your skin underneath the skin Which is nowadays if you go out to the commercial market buy a product like this, you have to buy something like that Which is commercially available today Of course, there's some in comfort people don't like putting something under the skin and this one is actually No pain and also not penetrable. It doesn't have to be put under the screen It's just patch it on the arm able to send all the signal wirelessly to a smartphone And they're able to capture that and do further analysis And there are the company doing this also this company are sugar beet no field sense or Google wise Graph where these are all start-up company trying to commercialize something like this in the market and Personally, I'm very excited about this because this is actually what a lot of people want Able to measure all this glucose level and making sure they are Monitoring that on a regular basis and able to adjust their daily behavior accordingly if they're seeing some Some changes in the level that they don't like to see So next I want to show you a video basically showing you how this watch can measure blood pressure It's important how you wear this watch like I said you use the wrist measurement you get this adjusted it's pretty easy to adjust you just lift this up you can move it up and down and this moves and basically the airbag is on one side and Once you do that now The thing is when you put it in you literally need to kind of put your two fingers here So do you get the exact spot on where to where you watch to ensure? And then you make sure to push the airbag down correctly and you get a Relatively snug fit because if you don't get a snug fit It's not going to do your blood pressure reading in fact when you do the ECG reading You might have to move it up a little more because it asks for the watch to be on the wrist bone a little more So how do you do your blood pressure reading? Press the top button top right hand button. You can either go Slide horizontal or vertical now you press this measure And then it'll tell you that you need to kind of lift your hand up so that it is Diagonally across your upper body, but it's not pressed down and you kind of support Your wrist It has given me a reading systolic diastolic including the pulse So as I said, I have exact watch Here right now. So after this talk if you are interested to try it out stay behind and then we can we can do some Live demo, which is the great part about being physically here on site for the conference So another interesting innovation about using smart watch to monitor health of the patient or persons This is one example that was actually launched end of last year in China What they actually did is they recording the coughing sound about 15 seconds and then basically pass that sound over to The smartphone and then a smartphone basically using certain AI model combining the data with the body temperatures with the breathing rate heart rate and blood Oxygen level and then use also data and pass it to the AI Engine and then doing inference and then basically find out if the person actually have a certain potential lung infection risk So this is very interesting And I foresee of course today in the cover situation. This might be a very helpful way to monitor covert patient at home In the last last point I actually want to raise up is actually The world long days coming up has September 25th So try to spread awareness about this And then the next topic I'll actually want to go into Deeper into looking at where the market is heading With last year Facebook changed their name to matter their lot of attention in the virtual world Metaverse and people started thinking okay with your virtual world Is there anything that can benefit the healthcare industries? Is there anything that can that the healthcare industry can make use of and then benefit the future development? So there are three areas. I believe it's actually very interesting One is actually medical device development They can actually try to use that Digital twin idea to improve the design development testing of these devices And then the digital twin of the human person They can actually use the generic generic makeup of the person that behavior the lifestyle of that person and then Duplicate as a digital twin and then basically apply a personalized treatment plan And that will basically help a fine-tune a personalized Recovery plan for those particular patient and then for the healthcare facility Is testing if the facility have a bottleneck in the future if they scale up to many patient coming here Which is highly likely the case in the early stage of COVID we all seen Sudden search of all these patient coming in can the hospital actually able to cater for that search What are other areas can improve in terms of optimizing the operation strategies and so on so so with that Let me show you a quick video actually done by Analytics insight which actually give a little bit more detail about the usage Digital twins are a digital representation of a physical object process or service It's key value proposition to digital twins is in their ability to combine real-time data physical dependency models and Intelligence from different platforms to simulate predict and improve assets and processes The healthcare industry is turning towards digital twins to improve customized medicine healthcare organization performance and new medicines and devices Top ways in which digital twins can transform the healthcare industry Digital twin can predict multiple outcomes for specific procedure of any treatment by simulating an invasive clinical procedure Digital twin can provide assistance in evaluating the right therapy and Diagnoses for each patient to cure chronic diseases by reducing medical risks Digital twin are used in hospitals to simulate workflow processes to detect potential errors in the existing system Smart wearables of patients can feed the digital twin in the cloud with real-time health data to detect certain symptoms at early stages Digital twin can enhance the performance of healthcare devices by running hundreds of simulations with different conditions in different patients Digital twin can create a digital copy of real patients to help in drug development and dosage optimization efficiently Doctors can receive sufficient data from patients body and observe different vitals medical conditions response to medicines diet blood sugar data and many more in the digital twin copy Digital twin can assist residents in medical training and diagnostics of patients copy for virtual surgery and a better understanding of human body anatomy physiological and anatomical difference from person to person Those simulations have been there earlier, too. They have been vital in today's healthcare industry Digital twins are capable of creating useful models based on information collected from wearable devices and patient records to gather data Digital twins are spreading in diverse streams based on other technological advancements and This opens the doors for treating and diagnosing patients with the help of the latest technologies For more such updates subscribe to analytics insight So this company actually give out this three scenario in a little bit more detail And I welcome you to actually including a link to that video. There's a little bit more detail that you can actually look at I will share the PDF Anyone interested? Send me a email. I can share the entire deck to you. So you actually have all these links And then kind of summarizing a little bit of my own observations about remote patient monitoring Of course, it's safer because the patient is going to be staying home and the doctor basically remote diagnostic monitoring the health of this patient So it's actually getting More and more popular especially during our COVID and the second point is I Believe the continuous patient monitoring that actually is going to be very important and then with the smart watch devices affordable actually affordable pricing right now Long battery life. There's extended long battery life Functions in a lot of the top brand watches. It can actually last like two weeks Without charging and then the interface actually have been improving You're obviously have seen a lot of top-coded like like Apple just recently announced some new addition to their watch products So this is actually Innovations that actually help push the entire market forward and then early health risk of detections Especially the patients is going to be wearing the watch 24 by 7 even sleeping so they can monitor the sleep pattern So that will help actually have the full picture of the entire day from morning to night Even sleeping time to detect that basically if that's potentially a health risk and then all these AI Advancements and today this morning. You probably heard the keynote Pytos was a leading framework for AI Development actually was now moved into leading foundation. So it's a Pytos foundation now So all these advancements in AI modeling, which what we also based on our research on is to analyze all this data from all the Patients and then able to come up with a meaningful analysis and then help the doctor to actually pinpoint the issues and the last point is open source Which obviously why we are here in this forum open source forum is the open source community There's a lot of related research Individual contributor contributing to the open source community on this a particular health care Kind of industry so that pushes all the tech companies all the research areas into advancing this area So we in this project. We're actually working with two researchers from University of Washington So dr. Erica Parson and also dr. Pierre Moritz And we are we are supporting them in their research Basically, they're using the MI ML model machine learning model to have a detection for the ischemic stroke patients and Right after this I'll pass it to my colleagues and he'll talk a bit more about how they're actually using some of the gizmo devices and AI model to analyze the patients and The last Talking about very high-level what we are trying to do here. So we have obviously the smartwatch on the left side Basically gathering all the sensor data. We take user privacy very highly Making sure all the user data being being protected There's several mechanism that we actually adopted of course the The HIPAA compliance and also we are talking about some of the latest advancement from tech company like Google So Google several months ago they announced this health connect technology and allowing Developers developing apps to access all these health data from the end user So there will be a user consent user have to consent Basically allowing developer to access certain of their health data and also the permission is not like all data Have to be specified with what kind of data developer are using and also they cannot Basically use the data for advertising purpose for example. So there's certain guidelines certain restriction being put together Allowing a central location storing all these health data being protected And not being misused by all these data So, you know, you know in our research basically all this data We're setting it to the smartphone and smartphone will basically have a feedback loop to allow the doctor to basically feedback to the patient At the same time the data will be basically sent over to the cloud to allow the doctors From the hospital able to see and analyze these data using AI model So we are using open source framework that we developed seven years ago It's called Cooper edge and there's an edge core and cloud core the edge core is actually hosted entirely in the Within the hospital private network. So it's totally protected. There's no outside connection to it It's only allow the doctors to analyze the data and then they're potentially by removing a lot of these users specific information The data to the all this data can actually be sent in a central location Allowing the AI model to be trained to be a much more efficient and much much better outcome Without particular user data. So still again, the user privacy is very important And then with that, I'll play the video of my colleagues who basically going through some of these Detail architecture and also demo within this second. This is about 20 minutes Okay, with the introduction president gave about the the smarter watch I'm here to talk more about how we can utilize the smart watch in the in the medical recovery scenario In this picture as we can see apparently the it's only the right half of the plate is the food on the right side of the plate Is it has been eaten? This is a special kind of situation where a patient if the patient suffered from a effect called left neglect This happens when the right side of the brain is damaged during a stroke So if the patient suffered a stroke and right side of the brain is damaged Then it affects how much the patient that can see on the on the left side Even though there's no problem to the to the eyes the the information collected from the On the left side is not able to process well in the right brain okay, and the the project originated from University of Washington Basel and it's called a stroke rehabilitation and project so in their original setting there's so the patient will be Fit into an environment with EEG machine to collect the the brainwave and There's a small gizmo machine. That's going to be following a curve The patient will follow that the we'll try to follow this gizmo machine and When the machine goes to the left and right If the machine goes out of the face what the patient can see then there's some information special about that left the neglected Condition from the EEG and the from the eye tracking So all together the eye tracking machine here and the gizmo and the EEG are there to collect data based on from the brainwave from the eye movement And the ultimate goal is to be able to help with these Post-stroke rehabilitation by using machine learning based the detection and analysis. This is really the the from the UW Department, so they have two department working on this is a collaboration between the the the medical Hospital and the university from the electrical engineering and the computer science Okay then the way we come in to help with is that originally the professors from the university wants to to improve the Capability by providing some local computing and collect data and the data can be processed here and originally the thing the thinking is so far everything we are seeing is only within the perimeter of one hospital and Apparently, there's the resource the computing resources the storage size the power consumption Those are things are very limited and also This is only they can collect data who is in this hospital. So there's a lot of a limitation to this setup. So the idea is to Provide some additional saying to provide some additional ways to help to improve the Capability when the first way to do this is by using a smart watch so the EEG machine and eye tracking are only two source of data to to work on this but currently as the president has mentioned the Smart watch currently can provide a lot more information like heart rate of blood temperature blood oxygen and also also force they can Not only be used in the hospital, but also they can use in a home environment. So it's very easy access for patients to use one example I can give is One of the professors is trying to use this heart rate Together with the EEG to to use the correlation to get rid of some noise collected during the EEG scanning So that's the first benefit of by using a smart watch that more Data will be provided more information with will be obtained and together with the EEG machine and eye tracking They can provide a lot of insight The second side that we want to improve this scenario is to provide the cloud computing and Edge computing all together the cloud computing and edge computing are to To achieve this feather is a learning It's a special kind of machine learning where you take models from different small models And that you aggregate them you get the insight among all those model and produce a more powerful machine learning model In addition to this the cloud apparently has more computing resources and has more storage and has more power that's way beyond what the hospital can provide and In the edge computing scenario It can provide a more variant essentially because as we said the for one hospital They can only have a certain amount of patients but if we combine the information collected from different hospitals then a more powerful a more insightful Machine learning model can be produced. So these are the three benefit the three key benefits So the smart watch provide more data the edge computing and the cloud computing all together will provide the more more computing power and analytical capability into the data collected and Speaking of these two benefits the there's also a question about the The data privacy currently I could go back here all the if we collect data from a hospital it's It's not a straightforward that we can send the data to the cloud for competition the other data in in the hospital has to be pre-processed to be filtered to be all those privacy information has to be Removed before we are able to send the data we have confidence to send the data to the cloud so in in our project, these are a few synergy ways that we are trying to provide the one is by using the edge computing we can Offer an opportunity to to pre-process the data before it goes out of the hospital and the second one of course is the capability of The cloud and the capability of the edge infrastructure themselves and the third one is we want to be able to provide a simple way For the hospital or for the patient to use the all these things for example for the cloud and for the edge We're using the open source Kubernetes and the cool batch Those are things that are great But they are not a straightforward to set up together We're together with for example the cloud and all these things for people who are not in with who do not have a computer background or even the cloud edge computing background It's not easy for them to set up And one of the goal for us is to have the doctor for example in the hospital To be able to plug this in to the power plug this into the internet and then they can start the collecting data So here I'm going to give a demo the demo for simplicity due to the time constraint We're going to show a simple end-to-end data flow the point is to show that we are able to collect data from a smart watch and the data will be collected by a Computer resources on the edge the edge is for for example hospital and eventually the data will be sent it to the cloud and The for further analysis so the data started from the watch is a hard rate and then it goes to the hospital It goes into the computing resources on the edge in the hospital and then the data will be eventually sent to the cloud if we just send this part as if we write this part separately is It's it's not as hard, but we here we want to utilize the powerful tools For the edge community for example corporate that is to manage all this right for you for all the hospitals Remember one of the goal is to make things easier and we do not want the hospital to have their own program running on their local computer then Then all the hospital do have different versions then once we come to deployment upgrade Those things gets more and more complicated and people eventually abandon them. We want to provide a simplified and Unified framework so that the hospital they can run the same program if we have update It's a very easy to out to update all this program either running in the cloud or running in the hospital and By saying that apparently there will be application running on the edge there will be application running in the cloud and to be able to Want to control this app to do the deployment to do the upgrade probably you already sense there's a corporate net is coming and In this demo, we're gonna put them in the Docker container and For the cloud side that we're using the Kubernetes and for the edge side that we're using the open source cool edge to To essentially add this node as one of the node in the cluster in the cloud We also have apparently this application is running on a worker node in the corporate net is cluster in the cloud But this is also running as an edge node as part of this class at the same cluster So the control plan is in the cloud. We have one Cloud worker node and we have one edge worker node So let's see the demo so for the to to simulate this edge Node so the calling at the edge means it's not part of the cloud. It has to be outside. What do we use here? So what we use here is AWS environment so to have something outside the AWS environment we put this node in We are using Google Cloud a Virtual machine from Google Cloud the goal here again is to separate to to say we have two nodes in two separate network One is the cloud one is some other network. This doesn't have to be in the cloud this can be a Raspberry by Pershing running in the hospital or some kind of machine running in the hospital So this is the reason we use Google Cloud is just so that it's outside the AWS network So here we can see on the left that we have the Kubernetes running on this two virtual machine This is the the master machine. This is the worker node And on the right is the Google Cloud edge machine so if we go to this See if we go to this Kubernetes control plan We can see that the three nodes that are running here. This is a control plan This is so the cloud the cloud master node. This is the see This is the cloud worker node and this is this one is the edge worker node Okay, currently we have two apps running One is the cloud and the cloud app and the other is the edge app the two app here If we go back to this picture is the cloud app here running in the cloud and the edge app here running on the edge All right. So that's the setup. That's the environment But okay for the watch we're for simplicity we're just using the this emulator and For this smart watch emulator We are using the heart rate The goal here is to whenever we change the heart rate on this smart watch Then it's going to send the data to the to the edge machine if we remember here again, so the watch will send data to the edge machine and This can be done either through Bluetooth or HTTP request here We're using HTTP request. It's also going to send the patient information That's the for us to say the watch usually carries some information that should not be sent to the cloud and we'll see that That information will be stripped by the edge machine before it sends to the cloud. Okay So here Is the edge machine? So this is the Google cloud as you can see from the Google So this is the the edge machine on this edge machine We're running some code to take data from the from the phone from the smart watch Just a code that we don't have to go into details here But the one thing I want to point out is When we receive data and the one we are about to send it to the cloud So actually when we receive data we are going to receive the patient information from the watch and Even though the patient information will be stored in the database But when we are preparing the data to be sent to the cloud as you can see we are only sending the heart rate and So a time we are not sending the patient this is to just to demo that the sensitive information from the smart watch can be filtered out that can be taken out of the data before it goes out of the hospital and Okay, so that's that the data. We are going to receive is in this file. So let's just Follow this file Here we already have some some data there. So that's fine. Let's do a follow here. Okay back to the watch Here again, we want to change the heart rate But before that, let's go inside the app and Start this app So here is you can see that this is the the IP address of this edge machine actually the part of the running the application running on the edge machine and If we start It was going to try whenever there's new data coming in. This is the code That's that so whenever the heart rate to change it will send the heart rate to this Application running on the edge. So this is the heart rate On this emulator. We can use this to let's change Let's just make a few changes to 500 which is super high for human And then back to 10 and then back to two nine seven. Okay Now let's get back to Here as we can see we have 500 297 so that means the data has been received on the edge machine if we use this The the data has come from Step one to step two right and step two of course will all continue we will get rid of the sensitivity information for the transmission to the cloud and It will post that the request it to the cloud. So let's back go back to the cloud machine See here we are going to the control plan This is the application running up in the cloud in container in the cloud so here Let's see we also have this file and as you can see it's getting there the five hundred five three one one 297 those data are just Coming in from the cloud if we go back to the picture here that the steps are We get data from the smart watch and the smart watch data will be either Transmitted through a Bluetooth to the edge the node or through HDV request These are still actually within the hospital the watch can be in the host the watch actually should be in the hospital Once we have the data the second step is so we collect this We do some cleansing get rid of the sensitive information and then send the data to the cloud once the cloud the data is in the AWS S3 The cloud app will pick that up So the data we just saw that that has picked up from the cloud machine They actually did not come from the edge directly that come from the cloud storage because cloud storage has more space So and also it decouples the relationship between that the edge and the cloud So the edge can just simply send that to the cloud and it doesn't have to care whether the cloud actually receive it at that moment Once the data is in the cloud or the cloud machine the cloud application will pick it up for some future process So that ends the demo Apparently the demo is the simplified version Again the goal here is to show the end-to-end Data flow once we have the data pass going from the watch to the edge to the cloud Then there are a few things we could do one is actually this is part of the Research together with the Uda Basel professors We can now provide a different With a smart watch We are able to provide more source of data and We want to use the federated learning which means On in each hospital will train a local model and with the local data So we have a different hospital these are different hospitals or each hospital will have their own local model with their local data and all this local model will be aggregated in the cloud Once the data is transmitted to the cloud So that's the process we were trying to demo there So the data is in the cloud we have data from different hospital the training process in the cloud We will aggregate them we'll just put them together and Send out useful for the information From the separate hospital back to their hospital So now they they updated the model can provide more capability to this local model for analysis and of course I want to mention that all this additional information it can be provided by Some medical device but with the the the line of smart watch that we are providing It makes things easier and actually the professor is working with us to to try to use the data provided from this smart watch For the open source of course the this whole thing With the cloud and edge we want to make things straightforward We want to make things easier to use so open source is a very natural choice for us So Kubernetes and the cool edge are used together to facilitate this the second part of Future step is cool batch given that it allow a node to be connected into the cluster in the cloud There's some limitation that Still single node running in the hospital if the hospital has more computing power For example if the hospital has their own Kubernetes cluster then how do we use this? For this kind of model here We have we had a future way has have Another project edge computing project. So we call this for next the information is here So in this project that we're extending the cool batch into supporting multiple clusters instead of multiple nodes on the edge That means we have control plan and that provides a lot of benefit with Coming from this so also for for example the Hierarchical of the cluster can can be connected in different ways instead of just two layers we can have different layers and This make the federated federated learning More attractive because now we have more regional model and the regional model can be Sent to the cloud for a higher level aggregation We also have edge edge communication and the currently there's a serverless project being developed so that we can provide more capability there In our mind that the the future would look like this We have a hospital and the hospital have the All kind of machines including the smartwatch to provide data for analysis and also it could be in the patient house With the patient house, we cannot bring all the medical device in by the way smartwatch It can also provide the some useful information and data. So all this data generated will be Processed locally and then processing the cloud to provide the more a better a much better Machine learning model to for for analysis Okay, so that's all my Talk, I'll hand this back to president. Thank you So with this detail discussion of architecture or the technology and open source that we are using I hope you have a good understanding about this research and also I think this is a very first step We will continue to invest and develop this. This is open source. So welcome anyone interested to Join this project and contribute your effort Will you look at our next step is look at Leveraging the smartwatch devices into different sensing data and also at the same time Figuring out if there's actually better framework to protect the privacy of data For example, as I mentioned earlier using Google the health connect. Is that a better way to do it or HIPAA compliance? So all these are the things that we will look into in the future And then of course from the hospital perspective the AI analysis part that will be working with University of Washington Some of the professor work already kind of started So we will be focusing on those areas So with that as I mentioned three questions first apologized to the Online audience who is not present here. This is only for the people here in this room. So first question Which country fastest? Yep, good Correct answer and How fast are they growing? Yeah, more than 500% so I have three colors Which one black blue and? Pink black Thank you. You go all right second question Name two types of sensor data that you can possibly capture using a smartwatch anyone heart rate and This is blue. Oh, this is pink There you go, okay last question unfortunately only one color left This is the one I mentioned earlier what day is the world long day this year you remember What correct good memory Unfortunately, only pink So thank you that conclude the talk today Again, if you have interest contributing to this open source project, please contact us And if you want to try out this Air pump with the blood pressure monitoring of this smartwatch Stay here and you can try it out. Also Yeah question go ahead Yeah, send me a email Yep right here So my name Preston dot now at visual weather comes any email I'll respond back to you with a PDF file Because we are running out of time But if you have questions feel free to stay behind and I can answer all your questions