news research photography about

Computer Aided Diagnosis in Mammography

I am a PhD student working on Computer Aided Diagnosis (CAD) of breast cancer in mammography at the Diagnostic Image Analysis Group in The Netherlands.

Early detection and treatment of cancer substantially increases the likelihood of survival and therefore, large scale screening programs have been implemented in many countries, where women over a certain are invited for biennial examinations. Mammograms, X-ray images of the breast, are recorded during a screening, which are subsequently inspected by one or more experienced readers for signs indicating the presence of a tumour, which is time consuming and subject to human error.

Recent advances in computer vision and machine learning, in particular deep learning gave rise to systems performing on par or outperforming humans on cognitive tasks such as speech and object recognition. Computers are consistent, not subject to fatigue or cognitive biases and have the potential to learn from an immense amount of examples, more than any radiologist will encounter in his lifetime and therefore have high potential in aiding or replacing humans in this field.

Important challenges are fusion of information from different sources; when a radiologist inspects an image she or he does not just look at a small region, but takes context, asymmetries between breasts, temporal change and patient background into account.

Bio

I obtained my B.Sc. degree in Artificial Intelligence at the University of Groningen and my M.Sc. degree in the same field at the Intelligent Systems Lab at the University of Amsterdam (Cum Laude). I worked as a visiting research student at Keio University, Japan for 8 months and as a research assistant at the machine learning group of the National University of Singapore/ A*STAR Bioinformatics Insitute for 6 months and recently visited the computer science group at Johns Hopkins for 8 months. My research interests are machine learning, higher level vision and decision making with applications in medical image analysis. Please see my CV for more information.

Publications

A Survey on Deep Learning in Medical Image Analysis Geert Litjens, Thijs Kooi, Babak Ehteshami Bejnordi, Arnaud Setio, Francesco Ciompi, Mohsen Ghafoorian, Jeroen van der Laak, Bram van Ginneken, Clarisa Sanchez

Discriminating Solitary Cysts from Soft Tissue Lesions in Mammography using a Pretrained Deep Convolutional Neural Network. Thijs Kooi, Bram van Ginneken, Ard den Heeten and Nico Karssmeijer. Medical Physics, 2017

Large Scale Deep Learning for Computer Aided Detection of Mammographic Lesions.
Thijs Kooi, Geert Litjens, Bram van Ginneken, Albert Gubern-Merida, Clarisa Sanchez, Ritse Mann, Ard den Heeten and Nico Karssmeijer. Medical Image Analysis - 2017

Covered by Aunt Minnie and Applied Radiology

Conditional Random Field Modelling of Interactions Between Findings in Mammography
Thijs Kooi, Jan-Jurre Mordang, and Nico Karssemeijer. SPIE medical imaging 2017, Orlando, USA

Deep learning of Symmetrical Discrepancies for Computer Aided Detection of Mammographic Lesions
Thijs Kooi, and Nico Karssemeijer. SPIE medical imaging 2017, Orlando, USA

A comparison between a deep convolutional neural network and radiologists for classifying regions of interest in mammography
Thijs Kooi, Albert Gubern-Merida, Jan-Jurre Mordang, Ritse Mann, Ruud Pijnappel, Klaas Schuur, Ard den Heeten and Nico Karssemeijer. International Workshop on Digital Mammography Malmo, 2016

Boosting classification performance in computer aided diagnosis of breast masses in raw full-field digital mammography using processed and screen film images.
Thijs Kooi and Nico Karssmeijer. SPIE Medical Imaging, San Diego, USA, 2014

Invariant Features for Discriminating Cysts from Solid Lesions in Mammography
Thijs Kooi and Nico Karssmeijer. IWDM, Gifu, Japan, 2014

Colour Descriptors for Tracking in Spatial Augmented Reality
Thijs Kooi, Francois de Sorbier and Hideo Saito. ACCV Workshop on Detection and Tracking in Challenging Environments, Daejeon, Korea, 2012. Demo 1 Demo 2

Region Enhanced Neural Q-learning for Solving Model Based POMDP's
Marco Wiering and Thijs Kooi, International Joint Conference on Neural Networks, Barcelona, Spain, 2010.

Master's thesis