Karim Armanious

Deep Learning / Generative Adversarial Networks (GANs)

Generative Adversarial Networks (GANs) is a new branch of deep learning which was developed recently in 2015. It is being hailed by many field experts as “the next frontier in deep learning” due to its potential in unsupervised learning (learning without labeled data) and its ability to create a common framework for several applications with no hand-crafted loss functions.

In GANs, two networks are pitted against each other and trained jointly. The generator network
acts as a team of counterfeiters trying to generate fake data that resembles the input data without detection. On the other side, the discriminator network is analogous to the police, trying to detect the counterfeit data. Competition drives both networks to improve their methods and learn more about the features of the input data.

GANs have been recently applied successfully in several applications including unsupervised image translation, domain adaptation, image in-painting and semi-supervised classification.

Emphases are:

• Development and optimization of new GAN algorithms
• Application of GAN methods for generation, domain-adaptation, and analysis of medical images including MRI, CT, and PET
• Application of generative frameworks on autonomous driving and acoustic applications

Publikationen

2021

Armanious, K.; Abdulatif, S.; Shi, W.; Salian, S.; Küstner, T., Weiskopf, D., Hepp, T., Gatidis, S. & Yang, B.
Age-Net: An MRI-Based Iterative Framework for Brain Biological Age Estimation 
IEEE Transactions on Medical Imaging, 2021, doi: 10.1109/TMI.2021.3066857.

Abdulatif, S.; Armanious, K.; Sajeev, J.; Guirguis, K. & Yang, B.
Investigating Cross-Domain Losses for Speech Enhancement
Proceedings of the IEEE European Signal Processing Conference EUSIPCO, 2021 . ArXiv pre-print: 2010.10468.

Eskandar, G.; Braun, A.; Meinke, M.; Armanious, K. & Yang, B.
SLPC: a VRNN-based approach for stochastic lidar prediction and completion in autonomous driving
Proceedings of the IEEE European Signal Processing Conference EUSIPCO, 2021 . ArXiv pre-print: 2102.09883.

2020

Armanious, K.; Jiang, C.; Fischer, M.; Küstner, T.; Hepp , T. ; Nikolaou, K.; Gatidis, S. & Yang, B.
MedGAN: Medical image translation using GANs
Computerized Medical Imaging and Graphics, Vol. 79, 2020, doi: 10.1016/j.compmedimag.2019.101684.

Armanious, K.; Kumar, V.; Abdulatif, S.; Hepp, T.; Gatidis , S. & Yang, B.
ipA-MedGAN: Inpainting of Arbitrary Regions in Medical Imaging 
IEEE International Conference on Image Processing (ICIP), 2020, doi: 10.1109/ICIP40778.2020.9191207.

Armanious, K.; Tanwar, A.; Abdulatif, S.; Küstner, T.; Gatidis , S. & Yang, B.
Unsupervised Adversarial Correction of Rigid MR Motion Artifacts 
IEEE 17th International Symposium on Biomedical Imaging (ISBI), 2020 , doi: 10.1109/ISBI45749.2020.9098570.

Armanious, K.; Hepp, T.; Küstner, T.; Dittmann, H.; Nikolaou, N.; La Fougère, C.; Yang, B. &Gatidis , S.
Independent attenuation correction of whole body [ ¹⁸F]FDG-PET using a deep learning approach with Generative Adversarial Networks
European Journal of Nuclear Medicine and Molecular Imaging (EJNMMI) Research, 2020, doi: 10.1186/s13550-020-00644-y.

Abdulatif, S.; Armanious, K.; Guirguis, K.; Sajeev, J. & Yang, B.
AeGAN: Time-frequency speech denoising via generative adversarial networks 
Proceedings of the IEEE European Signal Processing Conference EUSIPCO, 2020 , doi: 10.23919/Eusipco47968.2020.9287606.

Armanious, K.; Abdulatif, S.; Bhaktharaguttu, A.; Küstner, K.; Hepp, T., Gatidis, S. & Yang, B.
Organ-Based Chronological Age  E  stimation  Based on 3D MRI Scans 
Proceedings of the IEEE European Signal Processing Conference EUSIPCO, 2020 , doi: 10.23919/Eusipco47968.2020.9287398.

Armanious, K.; Abdulatif, S.; Kumar, V.; Hepp, T.; Yang, B. & Gatidis, S.
Adversarial Inpainting of Arbitrary shapes in Brain MRI
Proceedings of the International Society for Magnetic Resonance in Medicine (ISMRM), 2020 .

Abdulatif, S.; Armanious, K.; Bhaktharaguttu, A.; Küstner, T.; Yang, B. & Gatidis, S.
Organ-based estimation of the chronological age based on 3D MRI scans
Proceedings of the International Society for Magnetic Resonance in Medicine (ISMRM), 2020 .

Hepp, T.; Armanious, K.; Tanwar, A.; Abdulatif, S.; Küstner, T.; Yang, B. & Gatidis, S.
MoCo Cycle-MedGAN: Unsupervised correction of rigid MR motion artifacts
Proceedings of the International Society for Magnetic Resonance in Medicine (ISMRM), 2020 .

2019