3D Human Pose Estimation and Practical Project to Learn by Following
Are you interested in AI but tired of always hearing the same lectures? This is a hands-on project that will teach you how to generate and estimate 3D poses from human images and videos, run real code, and build your own data from images.
Once in theory, once in practice 3D Human Pose Estimation!
If this is my story, please pay attention.
✅
Anyone who wants to learn a new field instead of always following the same image classification/object detection models
✅
A college student who is about to start his or her graduation project but has not yet decided on a topic
✅
Students who want to participate in AI competitions with novel topics and fields
✅
For those of you who are curious about what they study in AI-related graduate schools
✅
Anyone interested in 3D human pose generation technology
✅
Anyone who wants to learn new technologies in the field of artificial intelligence
3D Human Pose Estimation is one of the fields of artificial intelligence computer vision, and is a powerful artificial intelligence technology used in a wide range of fields such as VR/AR, robotics, image analysis, avatar creation, and the movie business.
AI Healthcare Services
Pose estimation for sports player play analysis
Pose estimation for AR/VR game services
Pose estimation based on real video sources
Pose estimation for image generation
In fact, the number of 3D-related papers submitted in the computer vision field is increasing every year, and many studies are being conducted worldwide. However, it is not easy to find materials to learn related technologies, so many people hesitate to learn 3D Human Pose Estimation.
In this lecture, we will learn about the theory of 3D Human Pose Estimation and practice generating 3D Human Pose from an image using actual code .
Building a golf swing video Building a Practical Model 🏌️♀️
This lecture is a condensed version of the knowledge I have accumulated over the past two years on a topic I have studied at the Graduate School of Artificial Intelligence.
The goal of this lecture is to build a model specialized for a specific domain (golf swing) by building a training dataset with golf swing videos obtained from YouTube and training a 3D Human Pose Estimator without any correct data.
We've lowered the barrier to entry
Students will be able to experience artificial intelligence technology in the field of 3D Human Pose Estimation, which was previously difficult to access and difficult to find materials for, through in-depth explanations of the concepts.
Follow along, experience, and learn.
Instead of a theory class that starts from the bottom, we aim to teach you by following along. You will run model training/evaluation code together and experience the process of creating 3D human poses for prepared images.
Our ultimate goal is to “build my own dataset”!
💡 After attending the lecture, you will be able to build a dataset and train a network. Therefore, you will be able to plan a service that provides 3D Human Pose for specific human activities such as yoga, squats, bowling, etc., in addition to the golf swing as an example!
Knowledge sharer who created this course Introducing “Collider” Kim Hyun-woo.
The follower who created this course is🥽
I will provide you with research and project experience based on my knowledge of computer science and artificial intelligence, experience in numerous deep learning/machine learning projects, competition awards, and graduate school research experience .
Graduated with a master's degree in artificial intelligence from Korea University
SCI(E) papers, international conference presentations
Hello, I am Kim Hyeon-woo, a copycat. Two years ago, I launched the course 'Create a GitHub Blog in One Day' on Inflearn, and many people loved my course.
At the time, since it was my first time making lectures, I started the lectures with a light heart, but there were many shortcomings in many ways. I think I did not realize the weight of the responsibility, so I have stopped giving lectures. This lecture that I am revealing to you now is a lecture that I have improved and minimized the shortcomings based on my experience.
While completing my master's degree in Artificial Intelligence at Korea University in two years, I published a paper on 3D Human Pose Estimation at ACCV2022 (Asian Conference on Computer Vision), the world's 19th largest conference in the field of artificial intelligence and computer vision, and it was also selected for an oral presentation.
I created this lecture after seeing many people feeling tired or giving up on learning because of the lack of accessibility to specific areas of artificial intelligence. I included my learnings and trial and error from the past two years in this lecture. I hope that through this lecture, more people will become interested in the field of 3D Human Pose Estimation and gain a new sense of accomplishment :)
Q&A 💬
Q. I don’t know anything about artificial intelligence. Is there anything I can gain from taking this course?
Yes! Through this course, you will be able to become more interested in artificial intelligence. You will also be able to create your own data and your own model at any time using the model you have built. (This is a lecture that anyone can take if they can use Python with prior knowledge!)
Q. Is there anything I need to prepare before attending the lecture?
You will need a GPU (graphics card) and Ubuntu OS.
Please be advised in advance that refunds are not possible due to the above issue.
Q. What level of content is covered in the class?
Before running the actual model training and evaluation code, we will look at the paper together, but rather than directly teaching you the deep theory, we will guide you through the prior knowledge you need to know. I suffered for a long time because I did not know any prior knowledge related to this field while researching. I will proceed by explaining in an easy way what you need to know and what the sentences in this paper say.
💾 Please check before taking the class
For the practical training, you will need Ubuntu OS, a program editor, and a graphics card (GPU) . In the lecture, we will use Visual Studio Code as the program editor.
We provide source code to students. We also provide links and other convenient means to provide additional materials as needed.
Python proficiency is required. If you have no experience using Python at all, we recommend learning Python before taking this course.
This course is recommended for those who have taken a basic artificial intelligence or deep learning course .
Please do not distribute the lectures and study materials without permission. These are study materials that I have put my effort, time, and passion into.
Recommended for these people
Who is this course right for?
For those who are tired of object recognition like Yolo
People who want to build data and pose estimation from their own videos, such as golf or yoga
If you are in a hurry to complete a college graduation project or a competition but can't decide on a topic
Need to know before starting?
Python
Basic understanding of artificial intelligence or deep learning
Ngoài nội dung chuyên sâu
Đó là bài giảng đã giúp tôi nảy ra nhiều ý tưởng khi cùng nhau luyện tập.
Đó là một bài giảng rất hữu ích vì tôi tìm thấy nó khi đang tìm kiếm một chủ đề mới.