2020 healthhub datathon

The proposed method employs a noise-tolerant adaptive loss that handles partially annotated labels. Moreover, the proposed reconstruction loss leverages prior knowledge of neuronal cell structures to reduce false segmentation near noisy labels.

Healthhub가 주최하고 서울대학교 치과병원 영상치의학과가 주관하는 2020 healthhub datathon대회에서 Track B부분에서 1등을 하였습니다.

Research has difficulty because dynamic data has artifacts like motion, breath in MRI reconstruction. To solve these problems, I applied through the additional network and adversarial learning. As a result, The performance of PSNR and SSIM improved when comparing basic and proposed networks

교내에서 주관하는 아이디어 제품 개발인 캡스톤디자인 경진대회에서 우수상을 했습니다.