基於深度學習之視覺辨識專論(英文授課)
Selected Topics in Visual Recognition using Deep Learning
| 節 | 週四 |
|---|---|
3 10:10–11:00 | 基於深度學習之視覺辨識專論(英文授課) EC114 3 節連堂 |
4 11:10–12:00 | |
N 12:20–13:10 |
* 根據陽明交大上課時間表所列
Computer vision aims to enable computers to see, understand, and interpret the world like human visual systems. Deep learning technologies are at the core of the current computer vision revolution. Large-scale annotated data and affordable GPU hardware jointly allow the training of deep learning models with hundreds of layers and millions of parameters, which greatly improve the performance of various machine vision applications and even initiate new vision applications. In the course, I will first introduce some deep learning technologies that are widely used in computer vision research, including deep neural networks, convolutional neural networks, and generative adversarial networks. Then, I will cover some important vision applications such as object recognition, detection, and segmentation, and the corresponding advanced deep learning algorithms.
1. Basic knowledge of linear algebra and calculus 2. Programming experience such as Python 3. Deep learning programming skills such as Pytorch, Keras, or TensorFlow
Four homework assignments 72% (=18% x 4) Final project 28%
| 週次 | 主題 |
|---|---|
| 第 1 週 | Introduction to Computer Vision |
| 第 2 週 | Conventional Machine Learning I: AdaBoost for Face Detection |
| 第 3 週 | Mid-Autumn Festival: No Lecture |
| 第 4 週 | Conventional Machine Learning II: Support Vector Machines for Pedestrian Detection |
| 第 5 週 | Deep Neural Networks and Convolutional Neural Networks |
| 第 6 週 | Representative CNN Architectures I: AlexNet, VGG-Net, GoogleNet, and ResNet |
| 第 7 週 | Representative CNN Architectures II: DenseNet and Generative Adversarial Learning: GAN, cGAN, and CycleGAN |
| 第 8 週 | Object detection I: R-CNN, Fast R-CNN, Faster-RCNN |
| 第 9 週 | Object detection II: YOLO, SSD, and FCOS |
| 第 10 週 | Semantic/Instance Segmentation |
| 第 11 週 | Segmentation with Few Training Data Annotations |
| 第 12 週 | Image Super-resolution |
| 第 13 週 | Image Style Transfer, Video Frame Interpolation, and Video Synthesis |
| 第 14 週 | 3D Point Cloud |
| 第 15 週 | Final Project Presentation I |
| 第 16 週 | Final Project Presentation II |
| 第 17 週 | Guest Lectures (tentative) |
Ian Goodfellow, Yoshua Bengio, and Aaron Courville, Deep Learning, MIT Press, 2016 Richard Szeliski, Computer Vision: Algorithms and Applications, Springer Verlag London, 2011.
- 地點
- EC118
- 時間
- Thursday 3:00 pm ~ 4:00 pm
- 聯絡方式
- Instructor and Email: Yen-Yu Lin 林彥宇 lin@cs.nctu.edu.tw TAs and Emails: Jimmy Yang 楊証琨 d08922002@ntu.edu.tw Chia-Yu Ho 何佳諭 mylifeai1116@gmail.com Yu-Lin Lu 陸玉霖 ulin010101@gmail.com