 Good afternoon everyone. I am Vandit from Mitspilani. I am going to represent specifically tech power assessments. Starting from sounds, images and video, we will go to a lot of applications from different interns. But I will start with some of the common things of this presentation like the purpose. In particular arts of tech power assessments, we are focusing on automated assessments of unexplored areas. Now, what are the unexplored areas? Means you go to some online content and there are some, you learn from that website. There are some assessments, there are some MCUs, let's say, there are some questions and you will be elated according to that. But unexplored areas, let's say sound, let's say image and videos, there are unexplored areas particularly for assessment. So, when we started our intern, it was our first task to explore applications. That, okay, we decided sound, image and video. So, what are the applications and what are the factors we need to consider while deciding these applications? So, first we need to choose, but first we need to see the ideology of the space suit or the e3 store, which was explained by Ayushi. And when we think that, we need to put this application as a course with assessment on ID Bombayx platform. And it should be some course with learning assessment, which is useful for the students. So, these are the areas and applications, you can say, which were decided by us. In sound technology we have piano, tabla, then in image we have drawing, painting, shaping, and then for video we have gymnastics and this diving. So, one thing which was in our mind was that we need to choose an application which is feasible for this 50 days. We can complete in 50 days because any application, we need to give it a final shape. If we choose some large application, if we are not able to complete in 50 days, it is useless. Project methodology, this is very common for all the projects. So, when we decide application, we started searching for exercises, what are the exercises that we can give to students and now what is our task as a developer? So, these are the exercises student, what will student upload, what will the process and how this results will be shown to students. So, with that when we become developer from this stage, we started exploring open source technology which we can use for specific to applications, then what pre-processing we need to do. After pre-processing only the student file is like eligible to be matched with the teacher's file because that is how we will do assessment, student file, teacher file matching and then score calculation and then response to students because we need to tell student that these are the things you did wrong in this part, you did right in this part and final score. Obviously, we did test, we did test and the success of project obviously depends on the parameter tuning and testing means there are a lot of applications, these are all research, I am not saying exactly but kind of research like things. So, there are a lot of things we need to decide parameters at some stages. So, this is how we did all these projects. The pre-processing that is talking about is mainly concerned with negating the effect of differences between the environment capture. For example, the teacher has played tabla one type of tabla, student has played another type of tabla. We want to know whether student has played tabla properly, the recording may be different. So, the student tabla noise has to match the teacher tabla noise, I always call sound as noise. So, they always get upset, there is no music, there is only noise for me. So, basically the amplitude they have to match before they start matching that. Similarly, there is a image drawing image that is uploaded, the student can upload a larger image, smaller image. So, I have to scale the image, I may have to rotate the image before I actually match. So, that is the pre-processing that is required in all size, that is to take care of any discrepancy between what the student has uploaded, the capturing environment with student has uploaded and the teacher is capturing environment. Noise reduction also, noise reduction also is part of their exercise. Okay, images some clean up, I think they may have done or may not have done. What if student plays? Better than the teacher. Better than the teacher, here assuming that the teacher file is the most ideal file according to means when you are matching. No, there is no better, okay, because better is a concept based on the human ear, correct, it is not better. What we are considering is how much does it match? So, even if it sounds better, okay, he will be feeling, okay, because it is not matching the teacher. We don't know what the interval is. We are matching the, okay, to answer that question, one of the things which No, no, no, no, because there is no interval. Intervals, from intervals you want to say that, I am playing piano, then Saree, Gamma, Prudin sir, that difference. Yes, so teacher plays perfectly, okay, and you are assuming it. No, no, teacher plays something, student has to replicate. Applicate? No, but student has to replicate. Teacher makes a mistake. Student has to replicate the mistake. There should be some suggestion for teacher that there is some mistake somewhere. Suggestion from student to teacher, right? Not student, the system itself. System itself. So, you should do some basic analysis at least. Okay. If I like Talat Memoot, I cannot say that Muhammad Rafi is horrible, you sing like Talat Memoot. Okay, this is music, this is arts, nothing like what is perfect, okay. And then, something should be perfect, right? You are singing? No, no, if we actually take the music of all of those which have been sung in Bollywood, you will come to know what is, how they are sung, right? That is out. That was discarded in the beginning only. Okay, this is only music. That is one way of doing it, but one way would be to actually study the sound. Study the sound, okay. So, at the end, do analysis on it. Yes, sir. So, what you are doing is, teacher file is there and student file is there. So, student is uploading, it is comparing with teacher. You are saying that it should be, it should give analysis of teacher file also, right? Right. So, for analysis of teacher's file, that should be... Okay, okay. Another question. Yes, sir. A student is playing. Okay, sir. Okay. The student is actually not matching with the teacher. Student? Yes. So, do you give any suggestions? Yes. You will show. That was what I want to see. Yes, sir. So, I have just one comment. You are saying that teacher's file should be analyzed, right? Yes. So, at the end, that is some reference. Machine knows the music, right? Machine doesn't know the music. So, you have to write certain models, no? How machine will know the music? Yes. You are saying that you need to analyze the teacher's file also. Yes. Analyze teacher's file means what? We are analyzing teacher's file, obviously. But for evolution, for assessment, we are taking, it is a reference. You go ahead. You go ahead. Yes. And then show a demo. Correct, sir. So, I will ask Prashant for the demo.