In this talk, I will give an overview of the recent research advances made in the Image and Video Understanding Lab (IVUL) at KAUST, with regards to single-object tracking and activity recognition. I will highlight the merits of using structured and robust appearance representations (e.g. sparse, group sparse, and low-rank) in tracking a generic object in video, using the particle filter framework. Furthermore, I will present the details of a new large-scale activity recognition/detection benchmark, called ActivityNet. Similar to how ImageNet revolutionized object recognition in images, ActivityNet is expected to enable the design of next-generation action recognition/detection methods.
AGENDA:
12:50pm – 1:00pm: Bienvenida y Refrigerio
1:00pm – 1:30pm: Advances in Robust Single-Object Tracking and Activity Recognition
1:40pm - 1:50pm: Preguntas, discusión y retroalimentación.
Fecha: 9 de marzo del 2015
Hora: 12:50 m - 1:50 p.m.
Lugar: Edificio ML Auditorio C
Más información de Bernard Ghanem AQUÍ
Profesores: Pablo Figueroa, Fernando De La Rosa, Marcela Hernández, José Tiberio Hernández, Vanessa Pérez