ECE 49595 - Undergraduate Computer Vision
Course Details
Lecture Hours: 3 Credits: 3
Counts as:
- EE Elective
- CMPE Selective
Experimental Course Offered:
Fall 2023, Fall 2024
Requisites:
ECE 20875, ECE 36800, ECE 30100, ECE 30200, MA 26100, MA 26500
Requisites by Topic:
Python, data structures, rudimentary signal processing including filtering and convolution, probability, multivariate calculus, linear algebra
Catalog Description:
An undergraduate introduction to computer vision. The course will cover both classical and deep-learning approaches to recogniting objects in images and activities in video. Topics to be covered include scene classification, object localization, segmentation, and activity recognition. The course will cover both the theory behind methods to solve these problems as well as how they are implemented in systems like PyTorch.
Required Text(s):
None.
Recommended Text(s):
None.
Lecture Outline:
| Week(s) | Lecture Topics |
|---|---|
| 1-2 | Forward and reverse mode automatic differentiation, and backpropagation |
| 3 | Stochastic processes as probabilistic programs |
| 4 | Model-based object recognition |
| 5 | Object recognition with deformable part models |
| 6 | Segmentation with normalized cuts |
| 7 | Activity recognition with HMMs |
| 8-9 | Scene classification with Alexnet, VGG-16, Inception, ResNet, and DenseNet |
| 10 | Object Localization with selective search, Fast RCNN, and Faster RCNN |
| 11 | Segmentation with Mask RCNN, object localization with SSD and yolo |
| 12 | 3D space/time deep learning approaches to activity recognition |
| 13 | 2D and 2+1D deep learning approaches to activity recognition |
| 14-15 | Current topics: image and video captioning, video question answering, vision-language navigation, image and video synthesis |