Contributions to Deep Learning for Life Science Applications - Université d'Angers Accéder directement au contenu
Hdr Année : 2023

Contributions to Deep Learning for Life Science Applications

Résumé

In this manuscript, I present my research accomplishments in various domains, focusing on the development of machine learning and deep learning algorithms for analyzing images and signals. I outline my expertise in creating unique databases that have contributed significantly to research success, such as low-cost seedling growth, the 3D models of natural rosebush plants, and the AgTech data challenge. Furthermore, I detail my involvement in multimodal student behavior monitoring and a multimodal speaker recognition database. My research has focused on contributing to machine learning and deep learning, specifically in texture-based feature extraction, developing deep learning algorithms, and overcoming challenges related to image annotation. I have explored shallow learning techniques for life science imaging, such as local binary patterns and wavelet scattering transform. In deep learning, I have developed convolutional neural network models for microscopic image analysis and MRI, as well as recurrent neural networks and long short-term memory networks for spatio-temporal images. Additionally, I have examined multimodal CNN models and devised novel techniques to tackle image annotation challenges. My work has revolved around advancing the field of machine learning and deep learning for examining and interpreting images and signals across various domains, such as life science imaging and biometric analysis. My future research directions include minimizing dependency on manual annotation and developing novel techniques on multimodal generative self-supervised learning to extract meaningful and high-quality features from multimodal data. Beyond developing new methodologies, my research pursuits aim to foster lifelong training initiatives, introduce new courses, and offer mini-projects and internships for young students. I intend to bridge the gap between academia and industry, promoting the exchange of ideas, resources, and expertise. Ultimately, my research plans strive to inspire and empower the next generation of researchers and innovators by fostering an environment of collaboration, education, and innovation.
Fichier principal
Vignette du fichier
HDR_ report_Pejman-min.pdf (6.11 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

tel-04219846 , version 1 (27-09-2023)

Licence

Paternité - Pas d'utilisation commerciale

Identifiants

  • HAL Id : tel-04219846 , version 1

Citer

Pejman Rasti. Contributions to Deep Learning for Life Science Applications. Computer Science [cs]. Université d'Angers, 2023. ⟨tel-04219846⟩
85 Consultations
39 Téléchargements

Partager

Gmail Facebook X LinkedIn More