ID 2451: Multi-task Segmentation models for Tissue Segmentation in Wound Images
Description:
This project focuses on developing a deep learning model that accurately segments tissues in a wound using a multitask learning approach. This targeted solution aims to streamline the tissue segmentation process, ultimately contributing to the SWODDYS project’s larger goals.
Tasks
- Literature Review: Conduct a deep literature review of multi-task image segmentation models. and assess their applicability to wound tissue segmentation.
- Model Development & Validation: Develop, train, and validate a deep learning model using the available dataset. Compare with current state of the art models for Tissue Segmentation in Wounds.
- Comprehensive Documentation: Document all project steps, underlying assumptions, and acquired knowledge in a clear, concise, and organized manner in an IEEE conference paper format
- Regular Progress Updates: Maintain consistent communication by providing regular progress updates to supervisors.
Requirements:
- Technical Expertise: Possess strong experience in Deep Learning and Computer Vision.
- Programming Proficiency: Demonstrate proficiency in Python programming and libraries like PyTorch, NumPy,and OpenCV. Experience with Git and Slurm preferred.
- Communication Skills: Maintain a good command of the English language for effective written and verbal communication.
Application
- Deadline for applications: 01.11.2024 10:00 am CEST
- Email me your CV and transcript of records.
- IMPORTANT! Examples of previous work.
- A simple explanation of a relevant publication from recent years related to the topic.
Supervisors