Researcher

Title

• Assistant Professor, Department of Computer Engineering and Software Engineering, Polytechnique Montréal

• Ph.D. in biomedical engineering, Polytechnique Montreal, Canada
• Master in biomedical engineering, Polytechnique Montreal, Canada
• Diploma in biomedical engineering, École Supérieure d’Ingénieurs de Beyrouth, Saint Joseph University, Lebanon

My research program focuses on computer vision (capture and analysis of human pose and motion for medical, multimedia and industrial applications) and medical imaging computing (computer-aided diagnosis). In the first theme, my team is seeking to democratize motion analysis, moving away from highly controlled environments and closer to the patient's home. In the second theme, we aim to build knowledge bases from massive sets of medical images to efficiently assist clinicians in their diagnosis. Through highly applicative projects, I work closely with clinicians (intensivists, pathologists, ophthalmologists, physiotherapists, surgeons and orthopedists), industrial partners and artists (circus and music).

• Medical image computing
• Computer vision
• Machine learning
• Computer aided diagnosis
• Human shape analysis
• Human motion analysis

Career Summary

After completing her PhD, Lama Séoud undertook a postdoctoral internship in industry at Diagnos Inc, which she later joined as a researcher. She worked on computer-assisted diagnosis of retinal pathologies. In 2016, she joined the Vision and Graphics Research team at the National Research Council of Canada (NRC) in Ottawa. She worked for three years on integrating deep learning into the acquisition and analysis of 3D data, particularly of human subjects in motion. Enriched by these industrial experiences, she returned to academia in 2020 as a professor in the Department of Computer Engineering and Software Engineering at Polytechnique Montréal, where she leads the VisionIC research laboratory.

Publications

1. Faure, G*; Dal Soglio, D; Patey, N; Oligny, L; Girard, S; Seoud, L. (2025). ProtoSSDML: Self-Supervised Deep Metric Learning for Prototypical Zero-shot Lesion Retrieval in Placenta Whole-Slide Images. Computers in Biology and Medicine.
2. Ceglia, A*; Facon, K; Begon, M; Seoud, L. (2025). Real-time, accurate, and open source upper-limb musculoskeletal analysis using a single RGBD camera—An exploratory hand-cycling study. Computers in Biology and Medicine. 184: 109434.
3. Lescarbeault, E*; Kunz, M; Morcos, MW; Seoud, L. (2025). TADA-SAE: Exploiting bilateral symmetry in learned texture representations for medical thermal imaging. Computers in Biology and Medicine.
4. Lescarbeault, E*; Seoud, L. (2025). Breast cancer detection from thermal images using asymmetries in learned texture vectors. International Symposium on Biomedical Imaging, USA.
5. Hubert, C*; Odic, N*; Gharib, S*; Noel, M*; Hajzargarbashi, A; Séoud, L. (2025). MuViH: multi-view RGBD dataset for robust hand gesture recognition in a collaborative robotic finishing platform. Robotics and Computer-Integrated Manufacturing 94: 102957.
6. Faure, G*; Dal Soglio, D; Séoud, L. (2024). Automatic inflammatory lesions detection in placenta wholeslide images using self-supervised learning. Annual conference of the International Federations of Placenta Associations, Canada.
7. Khani, M; Cheriet, F; Seoud, L; Debanné, P; Parent, S; Labelle, H. (2024). Changes in trunk appearance after scoliosis spine surgery: Two-year follow-up study. Spine deformity. 12: 1071-1077.
8. Arnold, C* ; Jouvet, P; Séoud, L. (2024). SwinFuSR : an image fusion inspired pipeline for RGB-guided thermal image super-resolution. IEEE Perception beyond the visible spectrum, Workshop of the Computer Vision and Pattern Recognition (CVPR) conference, USA.
9. Tranchon, A*; Kunz, M; Séoud, L. (2024). Preoperative MRI whole-vertebrae segmentation in patients with severe adolescent idiopathic scoliosis. International Symposium on Biomedical Imaging, Greece.
10. Caron, R*; Londono, I; Seoud, L; Villemure, I. (2023). Segmentation of trabecular bone microdamage in Xray microCT images using a two-step deep learning method. Journal of the Mechanical Behavior of Biomedical Materials. 137: 1-9.