UROP Proceedings 2020-21

School of Engineering Department of Computer Science and Engineering 115 Deep Video Super-Resolution Supervisor: CHEN Qifeng / CSE Student: PARK Chun Ho / MATH-CS Course: UROP1100, Fall Although the problem of super-resolution has been explored thoroughly, super resolution of polarization images is an interesting problem that has not been addressed yet. Polarization images has numerous industrial applications as it allows tasks such as damage detection and reflection removal, therefore developing a super-resolution model on polarization images would allow improved performance and novel extensions on performing these tasks. This project aims to test existing super-resolution models developed for conventional images to polarization image dataset, and find out which models yield reasonable performance on this specific type of data. For this study, the SRCNN model and the EDSR model have been tested, and the EDSR model has shown superior performance in terms of PSNR and SSIM scores. Deep Video Super-Resolution Supervisor: CHEN Qifeng / CSE Student: REN Xuanchi / COSC Course: UROP1100, Fall We present a self-supervised approach with pose perceptual loss for automatic dance video generation. Our method can produce a realistic dance video that conforms to the beats and rhymes of given music. To achieve this, we firstly generate a human skeleton sequence from music and then apply the learned poseto-appearance mapping to generate the final video. In the stage of generating skeleton sequences, we utilize two discriminators to capture different aspects of the sequence and propose a novel pose perceptual loss to produce natural dances. Besides, we also provide a new cross-modal evaluation to evaluate the dance quality, which is able to estimate the similarity between two modalities (music and dance). Finally, experimental qualitative and quantitative results demonstrate that our dance video synthesis approach produces realistic and diverse results. Source code and data are available at https://github.com/xrenaa/Music-Dance-VideoSynthesis.