How DeepSORT Works?

Показать описание

So in this video, Im going to give to you a clear and simple explanation on how Deep SORT works and why its so amazing compared to other models in this computer vision lecture. We implement this in OpenCV in the following course:

⭐Want to learn cutting edge AI? Check out Augmented AI University👇🏼

But to understand how DeepSORT works, we first have to go back, waaay back and understand the fundamentals of object tracking and the key innovations that had to happen along the way, for DeepSORT to emerge.

Now tracking assumes that we have an already detected an object of interest. For detection as you may already know is done with YOLOv4. Once we have detected the object, it is assigned an id and is tracked using Deep SORT. We use an example with Elon Musk and SpaceX

So in this application, we have applied deep sort for tracking of vehicles on a highway for traffic surveillance applications.
------------------------------------------------------------
Learn Advanced Tutorials
Support us on Patreon
Chat to us on Discord
Interact with us on Facebook
Check my latest work on Instagram
------------------------------------------------------------

0:00 - Introduction
0:42 - Object Tracking
2:39 - Optical Flow and Mean Shift
3:01 - Mean Shift
3:58 - Optical Flow
5:10 - Kalman Filter
7:00 - Simple Online Realtime Tracking (SORT)
7:55 - Detection
8:21 - Estimation
9:00 - Target Association
9:31 - Track Identities life Cycle
10:30 - DeepSORT
11:27 - Deep Learning
12:13 - The Appearance feature Vector
13:03 - Climax of the Story

Рекомендации по теме

Комментарии

best video on deepSORT till date. very simple explanation. Thank You so much. I just started my computer vision journey. love from india.

ParthivShah

Phenomenal explanation that helped me understanding MOT in a single video after pondering for a month

brianyoon

Okay, so in this video I did a project by direct order from Elon Musk, there were rockets launched, Kanye West memes and an exceptionally great storyline accompanied by beautifully detailed animations.
The technical content was brilliantly conveyed, I was exceptionally entertained and took away more about the topic than after two weeks of literature research.

You, sir, have single-handedly beaten the entire educational system!
I love your content from the great humor - over the timestamps in the description - up to the immense quality of the videos!
Much love from Germany!

michaelschulte

I have been working on research in my university about tracking people for five months and I have watched many videos on the topic, but this video summarized everything wonderfully. Thank you from the bottom of my heart.
Can I get the images, video slides, or the name of the application that the graphics are working on in order to present the project at the university?

abdktefane

This is the best learning video I have seen on youtube. What an impressive way to explain DeepSORT, It cleared my doubts regarding Kalman filter. The musk story you make holds us till the last second of video. you are funny and smart! Thanks for this video :)

priyankasaini

Amazing video, thanks for the knowledge.

itaynave

OMG...there is so much to digest from this video... I need to watch this video at least 10 times to grab all...

ae

This video is really beneficial. Thanks a million!

aliabdari

Best video on object tracking !! 👍🏻❤️👍🏻

agrimarora

Nice video. What a cool way to learn something complex like Computer Science! Thank you <3

meomeo

It's so nice that you've read the SORT paper aloud 😰😰

ChetanAnnam

This is the most intuitive video on the topic !
Loved the teaching style

AbdulAhad-egse

Very good and funny overview, thanks :-)

papuce

Thanks for the video. I am curious for the sources for the fps of Tracktor++, JDE, and DeepSORT. E.g., how is tracktor++ 3 fps or how is JDE 12 pfs and low res (even on their official github they report 22.2 fps at 1088x608)? And when you say that Deep SORT is the fastest, do you consider the additional detection step?

katerynachumachenko