DISC: Latent Diffusion Models with Self-Distillation from Separated Conditions for Prostate Cancer Grading
Published in IEEE International Symposium on Biomedical Imaging (ISBI), 2024
Developed Latent Diffusion Models (LDMs) to generate high-fidelity histopathology images for training prostate cancer grading models. Introduced the DISC framework to enhance the accuracy of GG patterns from LDMs trained on imprecise annotations.
Extended version accepted to SynData4CV @ CVPR 2024.
Recommended citation: Man M. Ho, Elham Ghelichkhan, Yosep Chong, Yufei Zhou, Beatrice S. Knudsen, and Tolga Tasdizen. (2024). "DISC: Latent Diffusion Models with Self-Distillation from Separated Conditions for Prostate Cancer Grading." ISBI 2024.
Download Paper