Deep multi-task learning for image/video distortions identification

Zoubida Ameur (1, 2) , Sid Ahmed Fezza (3) , Wassim Hamidouche (1, 2)

1 INSA Rennes - Institut National des Sciences Appliquées - Rennes
2 IETR - Institut d'Électronique et des Technologies du numéRique
3 National Institute of Telecommunications and ICT

Identifying distortions in images and videos is important and useful in various visual applications, such as image quality enhancement and assessment techniques. Instead of applying them blindly, these techniques can be applied or adjusted depending on the type of distortion identified. In this paper, we propose a deep multi-task learning (MTL) model for identifying the types of distortion in both images and videos, considering both single and multiple distortions. The proposed MTL model is composed of one convolutional neural network (CNN) shared between all tasks and N parallel classifiers, where each classifier is dedicated to identify a type of distortion. The proposed architecture also allows to adjust the number of tasks according to the number of distortion types considered, making the solution scalable. The proposed method has been evaluated on natural scene images and laparoscopic videos databases, each presenting a rich set of distortions. The experimental results demonstrate that our model achieves the best performance among the state-of-the-art methods for both single and multiple distortions (Code is available at: Learning-for-Image-Video-Distortions-Identification).

Mots clés: Distortion; Identification; Multi-task learning; Deep learning; Natural image; Laparoscopic video; Multiple distortions; Scalability

Domaines: Sciences de l'ingénieur [physics]

Zoubida Ameur, Sid Ahmed Fezza, Wassim Hamidouche. Deep multi-task learning for image/video distortions identification. Neural Computing and Applications, 2022, 34 (24), pp.21607-21623. ⟨10.1007/s00521-021-06576-5⟩. ⟨hal-03464454⟩