Just brilliant Mitch! I like the fact that the system can detect and react appropriately to newly placed objects like the Kinect box. It's one thing for the network to prevent the cube from passing through a desk surface which occupies a set of XYZ coordinates that are discovered at the moment the program starts running but it's an even more impressive thing for the network to stop the cube from conflicting with the Kinect box whose presence, size and position were completely unknown to the network until it was moved into range at some random time long after the program had begun to run. How interesting that one neural network (viz. your mind) expressed frustration with another neural network (viz. the one embodied by your project) on those few occasions when the latter didn't operate in accordance with a desired function.
Hi Mitch, I'm in Grade 12 this year and I'm going to study Computer Engineering next year at UP. Your videos have been a great help. I was hoping you could maybe send me a list of software you used during your time there.
Its been great man,although you started way before I have watched your channel since move in day 😂😂when you had that Nasa Shirt on. 💪Now going into my Final year studying Mechanical Engineering, Me coming to UP in 2020 your videos played a role and I am glad I did 👍 I appreciate it man. All the best in your future endeavors...
hey can you share excel schedule time table.. awesome final year project, this is truly inspiring for me , someone whose a final year project in computer science....
Wake up 6 and go to bed at 00:00. Mostly work in there with breaks at lunchtime and in the evening for gym and dinner. That was kinda the schedule this year, obviously changed a lot throughout the degree
This is both exciting and terrifying. I'm approaching my final year in electronics engineering and choosing image processing as one of my electives. It's cool to know that this is a project I could do but also intimidating to try to get this done
That is an awesome project dude! I wonder if you could increase performance by letting the camera just calculate the surface when it appears (like the box) completely, than catching 4 points per side on a x,y,z cordinate to span a plane in the room and clear the cache to get some performance back. But again this is awesome keep it up my man.
@@mitch7w If it works, that's all that matters. As an outsider, it's always easier to make assumptions, so you probably had your reasons for the choices you made along the way. I also understand that time constraints played a big part. I hope the hard work paid off for you. Keep up the good work.
Mitch i need your help I'm a first year student at UP for computer engineering and I'm experiencing some problems when registering. While choosing modules, Cos 110 states that "Requisites are not met for class, not enrolled "
Hi....So if fail a module Neh,and it blocks the next year's module and so on ...will there be a point where you will be attending 2 or 1 module for the whole year?
Hey Mitch, seems like I am a year late to the party, but just wanted to let you know I absolutely love your project! I am an undergrad just finishing an ML class at UC Berkeley and have been on the search for cool summer project ideas to improve my resume. My idea was to use a raspberry pi with a camera, mounted on top of a VR/AR headset in order to be able to render complex virtual objects (like Architecture 3D models) over open spaces (like an actual construction site). As it seems you have already gone through the struggle and gained some experience in that field, I was wondering if you have any advice on approaching such a project. Based on your experience, do you think training those cnns might be too computationally difficult for the raspberry's processor? Also, is there some minimum hardware specs you could recommend for the camera module - do you think having a moving setup (like in a headset) might create additional performance issues? Those might be difficult questions to answer, I know, I would appreciate your perspective though ;) Thanks for the video and wish you all the best!
Hey man! Good luck with the degree. I don't think you'll be able to train on the Pi but can probably do inference if you're smart. I trained on a laptop. Use a basic webcam for video input (cheap and good enough quality) and then I don't have much experience with VR but whatever you can for cheapest and has the best API is probably your best bet! DM me on Twitter if you wanna chat more💯
SO MUCH. Some of my mates wrote entire components of their Projects in C++ for efficiency and then had to modify them later and rewrote in Python just cause of the ease
Now you have a cube that's not exist in reality, how about taking further steps, for instance, moving the object (cube) in augment reality and the 'same object' (cube) will move in reality
or, at 5:28 instead of colliding with the surface put some stuff (e.g a RC car) and when you collide the cube with right side of the car, it moves horizontally to the left, collide with left side it moves to the right, back side it moves forward, front side it moves backward, and at the edge the car spin around itself.
![Lethal Company Music Video 🎵 "Follow The Orders" [VERSION A]](http://i.ytimg.com/vi/QeooZB4Z7RA/mqdefault.jpg)
Had no idea what he was saying the whole time but he really did a great job convincing me to take up computer engineering
DO IT
good luck lol
Thanks man! It's a tough road but worth it
If you don't get a distinction for this project. I'll be very surprised. Bright future ahead for you mate.
Got that A. Life is good. Thanks man!
i am watching this video , when i am finding topics for my final year projects related to VR. I found this realy amazing. great job Mitch
Congrats Mitch, super proud of you mate
Thanks hey, feels great!
Just brilliant Mitch! I like the fact that the system can detect and react appropriately to newly placed objects like the Kinect box. It's one thing for the network to prevent the cube from passing through a desk surface which occupies a set of XYZ coordinates that are discovered at the moment the program starts running but it's an even more impressive thing for the network to stop the cube from conflicting with the Kinect box whose presence, size and position were completely unknown to the network until it was moved into range at some random time long after the program had begun to run. How interesting that one neural network (viz. your mind) expressed frustration with another neural network (viz. the one embodied by your project) on those few occasions when the latter didn't operate in accordance with a desired function.
Thanks Pete! Turned out so nicely
Hi Mitch, I'm in Grade 12 this year and I'm going to study Computer Engineering next year at UP. Your videos have been a great help. I was hoping you could maybe send me a list of software you used during your time there.
Watch enough videos and you'll hear mention but python, c++, lateX
I'm studying computer programming and I don't know what we would do with you and your ideas.
Haha, very kind.
Its been great man,although you started way before I have watched your channel since move in day 😂😂when you had that Nasa Shirt on. 💪Now going into my Final year studying Mechanical Engineering, Me coming to UP in 2020 your videos played a role and I am glad I did 👍 I appreciate it man. All the best in your future endeavors...
Thanks so much for being such a longtime fan. Good luck with the end of yours!
This looks amazing, congrats man
Thanks so much hey!
hey can you share excel schedule time table.. awesome final year project, this is truly inspiring for me , someone whose a final year project in computer science....
Congrats, can you maybe share your time schedule for the engineering somewhere? Would be interesting to see.
Wake up 6 and go to bed at 00:00. Mostly work in there with breaks at lunchtime and in the evening for gym and dinner. That was kinda the schedule this year, obviously changed a lot throughout the degree
This is both exciting and terrifying. I'm approaching my final year in electronics engineering and choosing image processing as one of my electives. It's cool to know that this is a project I could do but also intimidating to try to get this done
You can do it! Day by day and subsystem by subsystem
Awesome !
Congrats bro, we made it ❤️
We did indeed❤️
Hi Mitch. I wanted to ask where you bought your website domain or where you host it on?
GoDaddy, Firebase.
That is an awesome project dude! I wonder if you could increase performance by letting the camera just calculate the surface when it appears (like the box) completely, than catching 4 points per side on a x,y,z cordinate to span a plane in the room and clear the cache to get some performance back. But again this is awesome keep it up my man.
Yeah lots of things could be changed but ah it worked and if I built it again I think I'd redesign a lot of stuff to make it more efficient...
@@mitch7w If it works, that's all that matters. As an outsider, it's always easier to make assumptions, so you probably had your reasons for the choices you made along the way. I also understand that time constraints played a big part. I hope the hard work paid off for you. Keep up the good work.
What happens if you use it in front of a mirror?
Literally struggling with my final project 😢
Shame it's tough hey! Don't feel shy to reach out to other final years - can all help each other a lot just by chatting about your problems
do mind sharing some resources that helped you during the project
3b1b NN videos, geeks for geeks intro to CNNs, PyKinect and Tensorfloe documentation
I'm guessing this project was more software focused than hardware? Were there any real hardware aspects?
Almost entirely software-based
Mitch i need your help
I'm a first year student at UP for computer engineering and I'm experiencing some problems when registering.
While choosing modules, Cos 110 states that "Requisites are not met for class, not enrolled "
contact the faculty/admin
Hey I'm a first year too and I have the same problem. They said you can only register for it during orientation week
@@SHDWMhofulet message the student administration for that program, they helped me
can i have the project for the sample??
Hi....So if fail a module Neh,and it blocks the next year's module and so on ...will there be a point where you will be attending 2 or 1 module for the whole year?
Yeah, if you don't get academically disqualified for not passing the minimum number of modules
Can you share the git link of your project?
Look on my GitHub
Hey Mitch, seems like I am a year late to the party, but just wanted to let you know I absolutely love your project! I am an undergrad just finishing an ML class at UC Berkeley and have been on the search for cool summer project ideas to improve my resume. My idea was to use a raspberry pi with a camera, mounted on top of a VR/AR headset in order to be able to render complex virtual objects (like Architecture 3D models) over open spaces (like an actual construction site). As it seems you have already gone through the struggle and gained some experience in that field, I was wondering if you have any advice on approaching such a project. Based on your experience, do you think training those cnns might be too computationally difficult for the raspberry's processor? Also, is there some minimum hardware specs you could recommend for the camera module - do you think having a moving setup (like in a headset) might create additional performance issues? Those might be difficult questions to answer, I know, I would appreciate your perspective though ;) Thanks for the video and wish you all the best!
Hey man! Good luck with the degree. I don't think you'll be able to train on the Pi but can probably do inference if you're smart. I trained on a laptop. Use a basic webcam for video input (cheap and good enough quality) and then I don't have much experience with VR but whatever you can for cheapest and has the best API is probably your best bet! DM me on Twitter if you wanna chat more💯
You just gotta love python
SO MUCH. Some of my mates wrote entire components of their Projects in C++ for efficiency and then had to modify them later and rewrote in Python just cause of the ease
Now you have a cube that's not exist in reality, how about taking further steps, for instance, moving the object (cube) in augment reality and the 'same object' (cube) will move in reality
or, at 5:28 instead of colliding with the surface put some stuff (e.g a RC car) and when you collide the cube with right side of the car, it moves horizontally to the left, collide with left side it moves to the right, back side it moves forward, front side it moves backward, and at the edge the car spin around itself.
This could do with Omni wheel
Yeah maybe one day
Hi, what are the specs of your laptop?
M1 Mac
thanks@@mitch7w
Can i get ur github id name
Is there a repo with the code?
Private one yeah ;)
Sorry G you had to have got a first honura in your degree