COMPSCI 670 | Schedule

Date	Lecture	Readings	Logistics
9/7	Lecture #1 (Subhransu): Introduction and logistics	Szeliski book, Chapter 1 The speed of processing in the human visual system, Thorpe et al., 1996	Setup Piazza, Moodle, Gradescope
Module 1: Image Formation and Representation
9/12	Lecture #2 (Subhransu): Radiometry	Surface reflectance estimation and natural illumination statistics, Dror et al., 2001. Chapter 3 of Forsyth and Ponce book NeRF	HW1 out (due 9/26)
9/14	Lecture #3 (Subhransu): Light and color	Szeliski book, Chapter 2 Basic Color Terms : Their Universality and Evolution, Berlin and Kay, 1969 (book, Wikipedia article) Beau Lotto's TED talk
9/16	Lecture #4 (Zezhou, Ashish): Python tutorial (Friday)	Collab page
9/19	Lecture #5 (Subhransu): Light and color; Image formation	Szeliski book, Chapter 2
9/21	Lecture #6 (Subhransu): Image formation	Szeliski book, Chapter 2 Wikipedia article on camera Light stages at USC
Module 2: Basic Image Processing
9/26	Lecture #7 (Subhransu): Modeling images	Texture synthesis by non-parametric sampling, Efros and Leung, 99 Non-local means denoising, Buades et al., 11 OpenAI DALL·E 2 and GPT3	HW2 out (due 10/10)
9/28	Lecture #8 (Subhransu): Linear filtering	Szeliski book, Chaper 3 Hybrid images Berkeley segmentation dataset
Module 3: Correspondence, Alignment, Geometry
10/3	Lecture #9 (Subhransu): Optical flow	Lucas-Kanade optical flow Middlebury stereo dataset Szeliski book, Chapter 9
10/5	Lecture #10 (Subhransu): Feature detection and matching	Szeliski book, Chapter 7.1
10/10	No class (Indigenous Peoples Day)
10/12	Lecture #11 (Subhransu): Image transformations and alignment	Szeliski book, Chapter 8.1
10/17	Lecture #12 (Subhransu): Applications of image alignment	Szeliski book, Chapter 8.2 & 11 VGG image search	HW3 out (due 11/7)
Module 4: Fundamentals of Neural Networks
10/19	Lecture #13 (Subhransu): Intro to recognition	Visual categorization with bag of keypoints, Csurka et al., 2004 Histogram of Oriented Gradients, Dalal & Triggs, 2005 Textons, the elements of texture perception, and their interactions, Julesz, 1981 Computer vision datasets (via Roboflow)
10/24	Lecture #14 (Subhransu, remote): Project feedback
10/26	Lecture #15 (Subhransu, remote): Project feedback
10/31	Lecture #16 (Subhransu): Linear models	Perceptron paper, McColloch & Pitts, 1943 Linear classifier, Wikipedia
11/2	Lecture #17 (Subhransu): Neural networks	LeNet, AlexNet ImageNet OpenAI Microscope (model visualizer) Pytorch CNN tutorial
11/7	Lecture #18 (Subhransu): Neural networks		HW4 out (due 11/30)
Module 5: Advanced Topics in Recognition
11/9	Lecture #19 (Subhransu): Transfer learning	CNNs off-the-shelf, Razavian et al., 2014 Deep Filter Banks, Cimpoi et al., 2015 Bilinear CNNs, Lin et al., 2015 Taskonomy, Zamir et al., 2018 Task2Vec, Achille et al., 2019
11/14	Lecture #20 (Subhransu): Transfer learning
11/16	Lecture #21 (Subhransu): Object detection	Dalal & Triggs, Poselets, Selective search R-CNN, Fast R-CNN, Faster R-CNN YOLO, SSD
11/21	Lecture #22 (Subhransu): Object detection; Image segmentation	Mask R-CNN, He et al., ICCV 2017 Fully convolutional networks, Long & Shelhamer et al., CVPR 15 UNet, Ronneberger et al., MICCAI 15 Panoptic FCN, Kirillov et al., CVPR 19
11/23	No class (Thanksgiving break)
11/28	Lecture #23 (Subhransu): Image generation	Inverting deep representations, Mahendran & Vedaldi, 2014 Texture synthesis & style transfer using CNNs, Gatys et al., 2015 Texture networks, Ulyanov et al., 2016 Visualizing deep texture representations, Lin & Maji, 2016 Generative adversarial networks, Goodfellow et al., 2020 (CACM version) Deep image prior, Ulyanov et al., 2017 Some theory on the deep image prior (DIP as a Gaussian process, Spectral bias of DIP)
11/30	Lecture #24 (Subhransu): Image generation
12/5	Lecture #25 (Subhransu): Unsupervised learning	Colorization - Zhang et al., 2016, Larsson et al., 2016 Context prediction - Doersch et al., 2015 Jigsaw puzzle - Noroozi & Favaro, 2016 Image rotation - Gidaris et al., 2018 Instance discrimination - Dosovitskiy et al., 2015, Xu et al., 2018, MOCO, SimCLR, BYOL 3D - PointContrast, Convex Decomposition Audio - Owens et al., 2016 Video correspondence Jabri et al., 2020, Future prediction Walker et al., 2016 Generative modeling - RepurposingGANs 2021, DatasetGAN 2021 Generative or Contrastive? Saha et al., 22
12/7	Lecture #26 (Subhransu): Unsupervised learning
12/12	Lecture #27 (Subhransu): 3D shape understanding	Multi-view CNN VoxNet, O-CNN PointNet, Deep Sets, PointNet++, SplatNet A deeper look at 3D classifiers, Su et al. 18
12/13	Poster presentations (CS 150/151)