ETRO VUB
About ETRO  |  News  |  Events  |  Vacancies  |  Contact  
Home Research Education Industry Publications About ETRO

Master theses

Current and past ideas and concepts for Master Theses.

Improved Diagnostics by Assessing the Micromorphology of Breast Calcifications via super-resolution

Subject

Breast cancer can present as either masses or microcalcifications or both, and detection of both types of lesions is important for breast cancer screening. Masses are low-contrast larger soft tissue lesions and microcalcifications (MCs) are high-contrast small calcium deposits. To assist radiologists, Computer Aided Detection and Diagnosis (CAD) systems have been developed in order to detect the location and boundaries of the region of interest (ROI) and label it as benign or malignant.
ETRO developed a CAD system with a low number of false negatives and false positive rate when analyzing 3D high resolution micro-Ct images. It demonstrated the evidence of the link between image features of MCs, malignancy and tumor type by exploiting MCs individual properties.

Kind of work

This thesis proposal focusses on super-resolution (SR) methods for improving the automatic analysis of MCs. SR methods allow to generate high resolution images from low resolution ones. Unlike upsampling methods such as interpolation, they restore spatial high frequencies and compensate artefacts such as blur or jaggy edges. Multiple Image SR techniques assume that many LR images of the same scene or object are available. As each image contain a slightly different version of the real scene, it is possible to merge this scattered visual information into one, enhanced image. When only one LR observation is available, the problem is reduced to a highly ill-posed inverse problem called Single Image SR (SISR). Recent attempts to use deep learning for super-resolution reconstruction have used supervised learning, which requires pairs of low- and high-resolution images for training. This limitation hinders more practical applications of super-resolution reconstruction. Recently unsupervised methods using Generative Adversarial Network (GAN) based SR approaches, have been proposed for real-world super-resolution.
The purpose of this thesis is to propose unsupervised SR approaches to improve both micro-Ct images image quality and the performance of the CAD system. The SR images must therefore contain the same relevant features as HR images, matching the prior of the CAD system.

Framework of the Thesis

[1] E. Papavasileiou et al., "Towards a CAD System for Breast Cancer Based on Individual Microcalcifications?," 2018 IEEE 20th International Conference on e-Health Networking, Applications and Services (Healthcom), Ostrava, 2018, pp. 1-5, doi: 10.1109/HealthCom.2018.8531134.
[2] Willekens, I., Van de Casteele, E., Buls, N. et al. High-resolution 3D micro-CT imaging of breast microcalcifications: a preliminary analysis. BMC Cancer 14, 9 (2014). https://doi.org/10.1186/1471-2407-14-9
[3] Nasrollahi, K., Moeslund, T.B. Super-resolution: a comprehensive survey. Machine Vision and Applications 25, 1423–1468 (2014). https://doi.org/10.1007/s00138-014-0623-4
[4] S. Anwar, S. Khan, N. Barnes, “A Deep Journey into Super-Resolution: A Survey”, ACM Comput. Surv., 2020, 53(3), https://doi.org/10.1145/3390462
[5] S. Chen et al., "Unsupervised Image Super-Resolution with an Indirect Supervised Path," 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA, 2020, pp. 1924-1933, doi: 10.1109/CVPRW50498.2020.00242.

Number of Students

1

Expected Student Profile

Required qualifications:
• Following an MSc in a field related to one or more of the following: computer science, Electrical engineering, Biomedical Engineering, Applied Computer Science - Digital Health.
• Experience with image processing and signal processing.
• Experience with machine learning and statistics.
• Strong programming skills (Python).
• Interest in developing state-of-the-art Machine Learning methods and conduct experiments.
• Ability to write scientific reports and communicate research results at conferences in English.

Promotors

Prof. Dr. Bart Jansen

+32 (0)2 629 1034

bjansen@etrovub.be

more info

Prof. Hichem Sahli

+32 (0)2 629 2916

hsahli@etrovub.be

more info

Supervisor

Miss Redona Brahimetaj

+32 (0)2 629 2930

rbrahime@etrovub.be

more info

- Contact person

- IRIS

- AVSP

- LAMI

- Contact person

- Thesis proposals

- ETRO Courses

- Contact person

- Spin-offs

- Know How

- Journals

- Conferences

- Books

- Vacancies

- News

- Events

- Press

Contact

ETRO Department

Tel: +32 2 629 29 30

©2022 • Vrije Universiteit Brussel • ETRO Dept. • Pleinlaan 2 • 1050 Brussels • Tel: +32 2 629 2930 (secretariat) • Fax: +32 2 629 2883 • WebmasterDisclaimer