Designing and developing deep learning-based applications for video analysis tasks. These activities include:- choosing and training appropriate deep neural networks with the best trade-off between accuracy and computational efficiency (Pytorch and Tensorflow frameworks)- simplifying and optimizing deep neural networks to run on board of devices with limited computational resources - integrating these models in C/C++ video analysis software applications

Web and artificial intelligence application development

Progettazione e realizzazione di un prototipo per il miglioramento delle prestazioni dei tradizionali algoritmi di analisi video, utilizzando approcci basati su tecnologie semantiche e deep learning, finalizzato al retrieval di informazioni.Durata in ore: 275.