 Hello, I'm Bharath Rajagapan and I'm responsible for helping develop SD strategy in market development for augmented virtual and mixed reality. You know, we are at a really interesting time in industry today. What we're seeing is a convergence of technologies, products, solutions and services that cuts across a range of industries from augmented virtual and mixed reality to autonomous driving, to unmanned aerial vehicles or drones, and a range of mobile robotic applications from consumers and industrial and commercial. And what all these applications have in common is the need for machines to be able to see. So-called machine vision or computer vision and that's the basis for allowing vehicles or devices to be able to move in the environments without the aid of human intervention. And so today we're about to show you a demonstration that helps our customers and partners develop solutions that can take advantage of machine and computer vision. So in particular, what I'm going to show you over here today is a technique called structure from motion or SFM. As the name implies, structure from motion takes the data coming from a video camera, analyzes every single frame by frame by frame, and based upon what's in the frame is able to extract structured data and motion data. That's really important as you'll see in a few minutes over here. So on the monitor you'll see two images. On the image to the right is data, live data taken from a camera. This is a single camera monocular camera that's placed in the rear of a vehicle and it also has a fish out lens as you can tell and a fish out lens as you know allows you to see 180 degrees field of view. Onto the left is a dynamic trajectory and three degree construction mapping. So let's go a little bit more detail of what's really going on over here. So as you look at the features as you're driving the car, it turns out that you don't have to look at every single pixel. What you really care about is pixels that change. You want to track those pixels that change and by being able to do that you don't have to take the entire set of information and all the data that you get from a camera but a reduced subset that's enough to be able to reconstruct the entire scene that's being shown in front of you. So in this scene over here, what you see are all these lines. These lines represent what's called optical flow. So every single frame it tracks those pixels that are important and that are critical to the scene. Those are called key points in key frames and with those key points and those key frames you attach to it a key descriptor for every single vector over here and that information that gets that into the reconstruction engine and in real time you can take that data and reconstruct the trajectory of the car which is shown by these colored squares and all the information around the car all the objects which is called the point cloud and by being able to do that you can take the full set of data to reduce data set and with very high accuracy and already have precision be able to reconstruct the trajectory as well as the 3D reconstruction. In fact, this is so good that has accuracy to less than five centimeters which is very, very powerful. So to sum it up, it turns out that with a single camera you can analyze all of the frame take all the data that is required in those frames and synthesize from that critical information so you can reconstruct the trajectory as well as the image around it. This is quite powerful. Now this case of course is for an automotive automotive type of application but this exact same technique can be applied to a drone it can be applied to a robot that you may have in your house like a robotic vacuum cleaner it can be used for AR VR headsets and eyewear anywhere where you have to sense the environment using a machine using a computer. So that's the essence of structural motion. So from SD point of view we not only provide great sensors as you can see over here to be able to capture the image but also terrific algorithms very precise algorithms that can take that data and create meaningful information from it. And with this customers and partners can take this little bit of information that's been extracted and then be able to run their applications and the solution to their services. So that is the essence of structural motion. So hopefully you found this interesting and I value it to you and thank you for your time.