Please enable JavaScript.
Coggle requires JavaScript to display documents.
Prelim ppt - Coggle Diagram
Prelim ppt
Summary statement
In this proposal we describe three aims for deep learning based histopathology assessment of the fibrotic tissue microenvironment, including the development of deep learning-based image analysis algorithms, the designs of an intelligent acquisition system with real-time analysis capability, and a histopathological study that can be carried out based on the developed algorithms and software.
We first study the use of deep learning-based techniques to develop solutions for the challenges in computational histopathology including image and visualization enhancement using image synthesis, and automatic disease detection in whole slide images relying on only slide-level labels.
We further incorporate the deep learning models for histopathological image analysis to microscopy automation, leading to the design of an opensource Python and Jave-based intelligent multimodal acquisition system for whole slide scanning.
With the acquisition system, we propose to characterize the collagen fiber organization in chronic pancreatitis and pancreatic cancer, aiming to find out the distinguishing collagen-based features in these two cases which are considered assembling each other in histopathology for the massive fibrosis involved the tissue microenvironment
The future directions not covered in this proposal include the development of deep leaning-based collagen fiber analysis algorithms, active learning models with self-supervised contrastive learning, the incorporation of these models into the run-time intelligent acquisition system, and characterization studies of other diseases models with fibrosis.
-
-
-
When I joined, and doing what
-
-