Publications
2025
-
Maisonnave, L.*, Haroun, K.*, Pégeot, T. (2025).
Exploiting Information Redundancy in Attention Maps for Extreme Quantization of Vision Transformers.
The IEEE/CVF International Conference on Computer Vision (Accepted)
-
Szczepanski, M., Poreba, M., Haroun, K. (2025).
Where Do Tokens Go? Understanding Pruning Behaviors in STEP at High Resolutions.
Springer-Nature Computer Science (Accepted)
-
Haroun, K., Allenet, T., Chehida, K. B., & Martinet, J. (2025).
Dynamic Hierarchical Token Merging for Vision Transformers.
In Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
- Volume 3: VISAPP, ISBN 978-989-758-728-3, ISSN 2184-4321, pages 677-684.
-
Proust, M., Poreba, M., Szczepanski, M., Haroun, K.. (2025).
STEP: SuperToken and Early-Pruning for Efficient Semantic Segmentation.
In Proceedings of the 20th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
- Volume 3: VISAPP, ISBN 978-989-758-728-3, ISSN 2184-4321, pages 50-61.
2024
-
Haroun, K., Martinet, J., Chehida, K. B., & Allenet, T. (2024, December).
Leveraging local similarity for token merging in Vision Transformers.
In International Conference on Neural Information Processing (pp. 286-300). Singapore: Springer Nature Singapore.
-
Proust, M., Poreba, M., Szczepanski, M., Haroun, K..
Optimising ViT for Edge Deployment: Hybrid Token Reduction for Efficient Semantic Segmentation.
In EEAI 2024-2nd European Conference on EDGE AI Technologies and Applications.
-
Fabre, W., Haroun, K., Lorrain, V., Lepecq, M., & Sicard, G.
From Near-Sensor to In-Sensor: A State-of-the-Art Review of Embedded AI Vision Systems.
Sensors, 24(16), 5446.
2021