Best Computer Vision YouTube Channels [2024]
Computer vision is a mature field of computer science, while also being an active area of continued research. As a result of the long history of the industry, and its relevance to many real world problems, there are many tutorials, guides, and explainers online from which you can learn all things computer vision, from the fundamentals to practical application.
In this guide, we are going to walk through five YouTube channels that actively cover, or have previously covered in detail, computer vision or concepts important to understanding computer vision. Without further ado, let’s get started!
Roboflow
The Roboflow YouTube channel features an extensive range of videos on computer vision. You can find guides on state of the art computer vision models from Florence-2 to Segment Anything to YOLO11.
In modeling videos, you can find guidance related to use, fine-tune, and evaluate computer vision models. For instance, in the YOLO11 tutorial video, you can learn how to train your own YOLO11 model with a custom dataset.
You can also find guides that cover computer vision techniques and problems, from dwell time analysis calculation to using vision to track players on a football field.
Looking for a place to get started? We recommend:
- How to Train a YOLO11 Model
- Dwell Time Analysis with Real-Time Streams
- Football AI Tutorial: From Basics to Advanced Stats with Python
3Blue1Brown
3Blue1Brown is a mathematics computer vision channel. While known for its explainers on topics like calculus, the channel has published several popular, detailed computer vision and machine learning guides.
The “Deep Learning” series walks through the foundations of deep learning. Much of the foundational knowledge covered in the series is directly applicable to computer vision.
We recommend starting with the “But what is a neural network? | Chapter 1, Deep learning” video and working your way through the whole series.
In addition, if you are interested in natural language processing, later tutorials in the “Deep Learning” series cover how LLMs work in depth.
Learn OpenCV
OpenCV is a framework commonly used in computer vision. It implements many foundational algorithms used for everything from edge detection to image contouring to cropping to video processing.
Learn OpenCV is run by the maintainers of OpenCV as an educational resource. On the channel, you can learn about the latest computer vision models. You can also learn how to perform specific tasks with OpenCV, from thresholding to reading and writing videos.
To learn the basics of OpenCV, we recommend the “Getting Started with OpenCV” tutorial series.
Computerphile
The Computerphile channel is run by the University of Nottingham. In each video, a teacher or multiple teachers will walk through a topic related to computer science.
Computerphile has a whole series of videos on topics related to computer vision and image processing. These span from foundational image processing algorithms like edge detection to the foundations of CLIP, a multimodal vision model.
We recommend exploring this Computerphile Computer Vision playlist on YouTube and working your way through the videos that interest you.
Stanford CS231n
While not a YouTube channel per se, the Stanford CS231n computer vision course is available on YouTube. We recommend this course for a more academic look at computer vision.
In the course, you will cover:
- The foundations of neural networks
- Classification
- Loss functions and optimizations
- CNNs and their architectures
- Training neural networks
- Deep learning software
- RNNs
- Object detection and segmentation
- And more…
You can see all 16 videos in the course in the Convolutional Neural Networks Lecture Collection on YouTube.
Conclusion
Roboflow, 3Blue1Brown, Learn OpenCV, Computerphile, and Stanford’s CS231 course are all excellent resources for learning computer vision from videos.
Each channel focuses on a different topic. For example, Roboflow is focused on covering the latest breakthroughs in computer vision, whereas 3Blue1Brown covers more foundations of deep learning. Computerphile covers the theory behind many image processing techniques, whereas Learn OpenCV shows how to apply those techniques in practice.
Stanford’s CS231 course, on the other hand, walks through the foundations of vision with a rigorous, academic lens.
If you are looking for written learning materials to accompany your learning, check out the Roboflow blog. We have over 800 blog posts on topics from training state of the art vision models to explainers on how common image processing techniques work.