caMicroscope

Mentors: Student: Shakib Khan

The aim of this project is to focus on patch-level classification tasks within the domain of whole slide tissue images. The approach involves utilizing pre-trained models as encoders and applying them to the task of patch-level classification. By leveraging pre-trained models, readily available on platforms like Hugging Face, aim to expedite the development process while maintaining high performance. The ultimate objective of the project is to integrate these pre-trained models into caMicroscope, providing users with a convenient tool for downloading and training models tailored to their specific tasks. Deliverables: - A documented method for training a classification model using a pre-trained or foundation model. - Implementation of this method as a set of software components. - Integration of the components with caMicroscope.

Mentors: Student: Qianqian Zhou

Develop a web application within caMicroscope to enable users to visually triage multiple slides, adding annotations such as quality scores or re-scanning flags. This project aims to enhance dataset quality by providing an intuitive means for rapid slide assessment.

Mentors: Student: Yuka Usui

Many pathologists use caMicroscope to annotate WSI cases in their daily work. How pathologists annotate the slide and the different behaviors that different pathologists work on the same slide are very important. Currently, caMicroscope shows some basic Synopsis and Annotations information in the Slide Information Details modal and the center position, zoom level of the user view.The purpose of this project is to track and collect logs of annotations made by pathologists on a slide, visualize these data on a dashboard(or in other clear ways), and can share the results in a report format. This enables pathologists to know how other pathologists annotate the same slide.

Mentors: Student: cgdogan

My project goal was to add DICOM file support to caMicroscope, a Whole Slide Image viewer. The viewer backend (OpenSlide) didn't support DICOM, so we needed a second backend. Surprisingly, before the coding period started, OpenSlide did get DICOM file support and this gave me the opportunity to tackle the bigger challenge of integrating BioFormats into caMicroscope. This would be extraordinary, as BioFormats is a Java library with no C wrappers available online. I started by refactoring caMicroscope code to use a single base Dockerfile which compiled OpenSlide. Then I made a Java wrapper and prepared its corresponding C caller. For this, I had chosen GraalVM C API, but once I found out about its level of performance and its requirement to achieve high code covergage first (for metadata collection), I rewrote it for Java JNI. To make it available to the two derivatives of our base image, I refactored this to be a separate project. We also needed to make this available to our Python image metadata server, so I wrote a wrapper in Python using cffi.

Mentors: Student: Tarun Jain

This project aims to add support for Multi-Channel Imaging in caMicroscope, a platform used for visualizing and analyzing medical imaging data. Currently, the platform only supports datasets with a spatial representation for up to three channels. This project aims to create a function that can take in numerical data in CSV format with an arbitrary number of channels and select three fields to assign to red, green, and blue channels to generate an image. This project will allow Researchers and Deep Learning partitioners to make sense of higher-dimensional data and improve their analysis of medical imaging data.