MaLMS

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

Learning to store and read the guidance frames for spatiotemporal object segmentation

Supervision: Prof. Dr. Nassir Navab, Dr. Seong Tae Kim, Ashkan Khakzar

Abstract

Object segmentation on spatiotemporal data has made significant progress in recent years. However, most of the works have been designed to segment the object in spatiotemporal data from a single annotation in the first frame. However, segmentation performance across the entire spatiotemporal data varies dramatically according to the selection of the guidance frame [1-2]. In this project, we will investigate a way to fully exploit the information by learning to store and read the guidance frames for longitudinal medical segmentation.

Requirements:

  • Good understanding of statistics and machine learning methods.
  • Very good programming skills in Python & TensorFlow? / PyTorch?

Location:

  • Garching

Literature

[1] S. W. Oh et al. "Video Object Segmentation using Space-Time Memory Networks," ICCV2019.

[2] B. A. Griffin et al. "BubbleNets: Learning to Select the Guidance Frame in Video Object Segmentation by Deep Sorting Frames," CVPR 2019.

If you are interested, please contact us via e-mail: seongtae.kim@tum.de or ashkan.khakzar@tum.de

ProjectForm
Title: Learning to store and read the guidance frames for spatiotemporal object segmentation
Abstract: Object segmentation on spatiotemporal data has made significant progress in recent years. However, most of the works have been designed to segment the object in spatiotemporal data from a single annotation in the first frame. However, segmentation performance across the entire spatiotemporal data varies dramatically according to the selection of the guidance frame [1-2]. In this project, we will investigate a way to fully exploit the information by learning to store and read the guidance frames for two challenging tasks (Semi-supervised video object segmentation and longitudinal medical segmentation).
Student:  
Director: Prof. Dr. Nassir Navab
Supervisor: Dr. Seong Tae Kim, Ashkan Khakzar
Type: IDP
Area: Segmentation, Machine Learning, Medical Imaging
Status: open
Start:  
Finish:  
Thesis (optional):  
Picture:  


Edit | Attach | Refresh | Diffs | More | Revision r1.4 - 29 Jul 2019 - 10:28 - SeongTaeKim