Cervical Cancer Cytology
Abstract
Weekly task progress summary of the cervical cancer research project, including:
- Short and long term research tasks on models, techniques and methods for cellular image analysis.
- Analyze and research prominent papers to identify solutions for the project.
- Compile notes, slides, paper links and related source code
Weekly Task Progress
| Deadline | Name | Progress | Description | Research Papers | Materials |
|---|---|---|---|---|---|
| Sep 16 - Nov 16, 2024 | Ensemble Learning for Cervical Cancer Cytology | (--) Researching ensemble learning methods for cervical cancer. (--) Compare the performance between individual and ensemble model. |
N/A | [Slides] [Code] | |
| Sep 23, 2024 | Methods for Preprocessing and Data Split for Multi-Classification | (--) Researching the preprocessing and data split techniques. (--) Implementation these methods on cervical cancer datasets. |
N/A | [Slides] | |
| Jan 06 - Jan 13, 2025 | Self-Supervised Model Learning | (--) Overview the Self-Supervised Learning topic. (--) Introduction to the outstanding model of Self-Supervised Learning: SimCLR. (--) Researching the aspects of SimCLR model: architecture, I-O data, loss function, fine-tune strategies, etc. |
[PP] | [Notes] | |
| Jan 13 - Jan 20, 2025 | Implementation SimCLRv1 Model (Part 1) | (--) Building and designing the SimCLR model: backbone, loss function, fine-tuning, augmentation techniques, etc. (--) Conduct quick experiments and evaluations on a small subset of the cervical cancer dataset. |
[PP01] [PP02] | [Slides] [Notes] [Code] | |
| Jan 20, 2025 - Current | Implementation SimCLRv1 Model (Part 2) | Progress | (--) Expand experiments on various strategies and approaches: transfer learning, fine-tuning. (--) Conduct detailed evaluations of the results and compare the effectiveness of different methods. |
[PP] | Pending |
| Mar 28 - Mar 31, 2025 | Overview of VLM (Vision Language Model) and its general application in cervical cancer | Progress | (--) What is VLM? (--) How do models works? (--) Training methods: contrastive learning, cross-modal learning. (--) Applications of VLMs in healthcare. |
[PP] | [Notes] |
Quick Links
Including quick links folder of project: resources and research paper.
Datasets
The cervical cancer dataset from Hospital A in Thai Nguyen includes images of both common and rare cancerous cells.
Account: BOCSDL@ai4med.com
Password: BenhvienAThaiNguyen