Dottorato in Scienze Psicologiche e Psichiatriche (ultimo ciclo attivato 28° - anno 2013)

Corso a esaurimento (attivi gli anni successivi al primo)

Insegnamento "Perceptual Image Modeling: bridging Vision"

Periodo
dal 17 novembre al 15 dicembre 2010, tutti i martedì dalle 10.30 alle 12.30 (aula B - Facoltà di Scienze) e dalle 14.30 alle 17.30 (laboratorio Gamma - Facoltà di Scienze)

Docenti
Gloria Menegaz

Cicli in cui è offerta

22° ciclo
22° ciclo
22° ciclo
22° ciclo
23° ciclo
23° ciclo
23° Ciclo
23° Ciclo
24° ciclo
24° ciclo
24° Ciclo - Neuroscienze
24° Ciclo - Scienze Psicologiche e Psichiatriche

Descrizione

Perceptual signal processing is about the investigation and modeling of the processes underlying human perception with the goal of providing an automatic system the sensory information which is available to humans. This opens the way to novel interaction modalities between humans and automatic systems by synchronizing them on a new generation of sensory-based communication channels. Among the different types of sensory information that are suitable for being exploited in such a framework, the focus here will be on vision and visual perception, in view of the generalization to a wider context accounting for other sensory inputs coming from different sources like audition (sound and speech processing), olfaction, tactile and eventually other types of behavioral signals. Perceptual image processing aims at designing models for vision and visual perception to support the representation and interpretation of the visual information within a digital system. In such a framework, vision models are defined in terms of a set of parameters that are relevant to perception. Understanding the visual system and designing models for artificial visions are two shades of the same object, and can only be accomplished on the ground of a unified perspective. The basic idea is to identify the essential visual cues that the visual system extracts and processes, for a given task, to derive a perception driven representation of the stimulus. The guideline is in the blending of the methodologies and know-how of information theory and signal processing at one side and vision sciences at the other for the development of a new man-centered approach to imaging. The lectures will focus on some perceptual mechanisms of different levels whose modeling strongly impacts on applications. Among these are visual stimulus encoding, post-receptoral mechanisms for texture perception, color vision and cognition, as well as their modeling and exploitation for image processing and medical imaging.

Allegati

Documenti