EECE 5642 Data Visualization Homework 2 Instructor: Mr. Yulun Zhang TA: Ms. Huixian Zhang, Room 427, Richards Hall, Email:
[email protected] Presentation Registration Due: 11:59 pm Feb. 9, 2020 by email Submission Due Date: 11:59 pm Mar. 16, 2020 by Blackboard Paper List [1] Anat Levin, Dani Lischinski, and Yair Weiss. Colorization using optimization. In ACM SIGGRAPH, 2004. [Paper] [Code] [2] Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Image super-resolution using deep convolutional networks. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2015. [Paper (conference)] [Paper (journal)] [Code] [3] Xun Huang and Serge Belongie. Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization. In IEEE International Conference on Computer Vision (ICCV), 2017. [Paper] [Code] [4] Jiahui Yu, Zhe Lin, Jimei Yang, Xiaohui Shen, and Xin Lu, Thomas Huang. Generative Image Inpainting with Contextual Attention. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018. [Paper] [Code] [Demo] [5] Yulun Zhang, Kunpeng Li, Kai Li, Bineng Zhong, and Yun Fu. Image Super-Resolution Using Very Deep Residual Channel Attention Networks. In European Conference on Computer Vision (ECCV), 2018. [Paper] [Code] [6] Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks. In IEEE International Conference on Computer Vision (ICCV), 2017. [Paper] [Code] [7] Yijun Li, Chen Fang, Jimei Yang, Zhaowen Wang, Xin Lu, and Ming-Hsuan Yang. Universal Style Transfer via Feature Transforms. In Advances in Neural Information Processing Systems (NeurIPS), 2017. [Paper] [Code] [8] Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-Image Translation with Conditional Adversarial Networks. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017. [Paper] [Code] [Demo] Paper Assignment We require 8 volunteer groups for presenting the above papers, where we discuss 2 papers per class. All the groups are encouraged to register for the presentation. For the group who presents the paper, they will get one extra bonus for each group member. The way of presentation registration is to send our TA an email with your group number, the order of your candidate paper list (like “1-2-3-4-5-6-7-8” meaning you want to select Ref [1] the most), and also the date you wish to present. First come, first served. The presentation assignment will be scheduled upon your registration sequence. The registration will be due on 11:59 pm Feb. 9. After that, if necessary, we will randomly pick up groups for paper presentation without giving bonus. Hint: one bonus point could add 5 pts to the final score for each student in the volunteer groups. Discussion Rules Every group who presents the paper needs to design at least 5 questions along with the answers for the discussion. Students should answer the question in class, and each question could have multiple answers; or ask one “relevant” question for the paper. ▪ Students who would not present are required to do such responses (questioning or answering) at least twice (15 pts x 2) and write a report (70 pts) independently to summarize one paper among the ones you did response. ▪ For the group who gave the presentation, they need to submit their slides (100 pts) for the assignment. Presentation and Report The presentation time is about 20-25 minutes. Each talk may have about 20-30 slides. The report length should be within 4 to 6 pages, by any standard report template (e.g., the NeurIPS template). Latex is recommended to edit your report (e.g., Overleaf). Both slides and report should cover the following contents: 1) What is the task/motivation in the paper? For example, the problem that the authors target to solve in their work. (20%) 2) What is the data they used in their work? How do they visualize the data in their method? (20%) 3) What are the key insights in their paper? Like novelty, analyses, observations, etc. (20%) 4) How did the author visualize their model/results? (20%) 5) What have you learned from this work? (20%) Each part above counts 20% of the score of slides/report. For report submission, it should highlight the questions you answered/asked in class.
