Lei Zhou, Ph.D. Research Proficiency Presentation: 'End-to-end orientation estimation from 2D cryo-EM images'

Tuesday, August 17, 2021 - 10:00am to 11:30am
Zoom - contact events@cs.stonybrook.edu for Zoom info.
Event Description: 

TitleLSAE: Chest Radiograph Disentanglement by Lung Swapping Auto-encoder for Thoracic Disease Interpretation

Chest radiographs (CXRs) are often the primary front-line diagnostic imaging modality. We observe that pulmonary diseases, such as COVID-19, manifest as characteristic changes in lung tissue texture rather than anatomical structure. Hence, we expect that studying changes in only lung tissue texture without the influence of possible structure variations would be advantageous for downstream prognostic and predictive modeling tasks. In this paper, we propose a generative framework, Lung Swapping Autoencoder (LSAE), that learns a factorized representation of a CXR, to disentangle the tissue texture representation from the anatomic structure representation. Upon learning the disentanglement, LSAE is leveraged in two applications. 1) After adapting the texture encoder in LSAE to the thoracic disease classification task on the large-scale ChestX-ray14 database, we achieve a competitive result with unsupervised pre-training. Moreover, when compared with Inception v3 on our multi-institutional COVID-19 dataset, COVOC, for a COVID-19 outcome prediction task (estimating need for ventilation), the texture encoder achieves less error with a smaller model size, further demonstrating the efficacy of texture representation for lung diseases. 2) We leverage the LSAE for data augmentation, by generating hybrid lung images with textures and labels from the COVOC training data and lung structures from ChestX-ray14. This further improves ventilation outcome prediction on COVOC.
Hosted By: 
Dimitris Samaras
Computed Event Type: