• Topic 1 (Clustering): K-means, probabilistic K-means, kernel K-means, entropy clustering, spectral clustering, bound optimization, biases.
• Topic 2 (Binary segmentation): explicit and implicit boundaries, deformable models, graph cuts, submodularity, interactive segmentation.
• Topic 3: Stereo and single-view reconstruction.
• Topic 4 Multi-valued labeling, regularization models, optimization algorithms, expansion, ICM, mean-field, annealing, message passing, relaxations.
• Topic 5 (guest lecture by Olga Veksler) Semantic segmentation: from shallow to deep.

May 30

• Meng Tang: Weakly-supervised segmentation and detection

July 11

• Zilong Zhong: "Devil Is in the Edges: Learning Semantic Boundaries From Noisy Annotations"
• Mengyun Zheng: "Deep Image Prior"
• Zhongling Wang: "Image Restoration without Clean Data using CNN"
• Kyle Robinson: "Image Style Transfer Using Convolutional Neural Networks"

July 18

• Paulo Pacheco: "Attention mechanisms for classification CNNs"
• Beracira Chen: "Detection of Long Edges on a Computational Budget: A Sublinear Approach"
• Zhixiang Meng: "Colorful Image Colorization"
• Ahmadreza Jeddi: "Unsupervised Visual Representation Learning by Context Prediction"
• Hsiu-Wei Yang: "Relation Networks for Object Detection"

July 25

• Dmitrii Marin: "Overview of Adversarial Learning in Vision"
• Andrew Na: "Few-shot adverserial learning of realistic talking head models"
• Gaurav Sahu: "Double-DIP: Unsupervised Image Decomposition via Coupled Deep-Image-Priors"
• Rex Zhang: "Seed, Expand and Constrain: Three Principles for Weakly-Supervised Image Segmentation"
• Zejun Yu: "Unsupervised monocular depth estimation with left-right consistency"