Home link News


Two presentations at MIDL 2018 [May 2018]

The first scientific results from our dementia research project together with University College London will be presented at the International conference on Medical Imaging with Deep Learning (MIDL) as an oral presentation on Wednesday July 4th at 10:20 and as a poster on Friday July 6th at 10:00.

The papers are already up on OpenReview.net:
- Simultaneous synthesis of FLAIR and segmentation of white matter hypointensities from T1 MRIs
- Robust training of recurrent neural networks to handle missing data for disease progression modeling

See you at MIDL!

Two PhD scholarships in machine learning for dementia diagnosis [September 2016]

Biomediq has two open PhD positions within machine learning applied to produce imaging biomarkers and disease progression models of dementia. Both positions are part of a larger research project in collaboration with University College London. Employment will be at Biomediq A/S in Denmark while enrolment will be at University College London in the UK from where the degree will be obtained.
The first project is about applying deep learning to produce imaging biomarkers of brain white matter pathology. For more information, see here.
The second project is about disease modelling of dementia using imaging biomarkers. For more information, see here.
The application deadline for both projects is October 23 2016 (midnight GMT).

Biomediq and UCL start new Dementia Research Project [August 2016]

Biomediq has started on a 4-year research project together with Nick Fox and Sebastien Ourselin from University College London on modelling vascular and white matter lesion pathology in dementia.

Biomediq wins the Knee Cartilage Segmentation Challenge! [June 2016]

At the 9th International Workshop on Osteoarthritis Imaging, Biomediq participated in the Knee Cartilage Segmentation Challenge (KCSC) with it’s Knee Imaging Quantification (KIQ) framework. The participants were provided training knee scans with segmentation masks for the cartilages and then had to submit segmentations for another set of validation scans. At the workshop KCSC presentation, KIQ was revealed to be the winners of KCSC. The official results with full details are pending.

Prestigious presentation at RSNA [July 2015]

Biomediq will present its results from the EU funded project ASSURE in the major lecture hall, the Arie Crown Theater with more than 4000 seats, at this years Radiolocal Society of North America meeting in Chicago, Il, US, Dec 2 11:10 AM. The results will illustrate the patented technology capable of identifying women of decreased mammographic sensitivity even though they are of low mammographic density. RSNA Webpage.

Breast texture presentations at UCSF workshop [May 2015]

Biomediq's co-workers from University of Copenhagen and Rigshospitalet presents Biomediq's technology on texture-based breast cancer risk assessment for screening stratification in two oral presentations at the 7th International Workshop on Breast Densitometry and Cancer Risk Assessment at UCSF June 10-12, 2015. UCSF Breast Density Workshop Webpage.

Biomediq publishes largest validation of automatic knee MRI segmentation [April 2015]

Biomediq published a validation of our fully automatic knee MRI segmentation framework in the Journal of Medical Imaging. The validation including more than 1700 knee MRI is the largest, published validation of automatic knee segmentation. The results demonstrated that our method matches or exceeds the performance of competing methods. The paper is available with open access.

Biomediq wins Best Presentation award at ECR [March 2015]

At the European Conference on Radiology, Biomediq won the award for the best oral presentation in the Neuro track for our Improved Alzheimer?s disease diagnostic performance using structural MRI: validation of the MRI combination biomarker that won the CADDementia challenge talk. The winners in the other areas are here.

Biomediq in the Danish news [November 2014]

Biomediq made the front page of the Danish newspaper BT with a feature story on early detection of Alzheimer's disease. A Danish description with link to news media mentions is at the homepage of the Department of Computer Science, University of Copenhagen.

Two oral presentations at RSNA [October 2014]

Biomediq will give two oral presentations at the 100th annual assembly of the Radiological Society of North America - the premier meeting for radiological sciences and technology (RSNA 2014).

Alzheimers Diagnosis and Hippocampal Volume

In the first abstract, we demonstrate that our state-of-the-art multi-atlas hippocampal segmentation technology based on the Harmonized Hippocampal Protocol outperforms FreeSurfer in terms of both diagnostic ability and clinical utility for diagnosis and monitoring of Alzheimer's disease.

Breast Cancer Risk

The second abstract demonstrates that by combining a representative range of previously published mammographic texture measures with a novel machine learning based approach, it is possible to separate future breast cancer cases from healthy controls to a degree that a) clearly improves on what is attainable through mammographic density alone and b) could facilitate personalized screening of, e.g., high risk women.

To set up a meeting at RSNA and hear more about Biomediq's pioneering imaging biomarker technologies, please contact Biomediq.

See you in Chicago!

Biomediq wins the MICCAI Dementia Grand Challenge! [September 2014]

At this years MICCAI conference held at MIT in Boston, Biomediq participated in the CADDementia grand challenge along with teams from the worlds leading groups in analysis of brain MRI. Among the 29 submitted entries, Biomediq won the challenge. Our methodology proved to be best for diagnosing the subjects into healthy, mild cognitively impaired, or Alzheimer's subjects. The framework analyzed T1 MRI and included fully automatic segmentation of selected brain structures and texture analysis of the hippocampi.

For more information, contact Biomediq or see the announcements at the University of Copenhagen and read the challenge details at Challenge website