Single chip “smart” cameras have recently begun to appear and reach the market. These devices are complete System-on-Chip [SoC] capable of acquiring images at high speed and performing all the computation required to give users a set of high-level informations about the scene. The attractiveness of these devices comes from their low-cost, small-size and high-speed characteristics, making them particularly suitable for consumer applications (such as providing supplemental informations to drivers when installed on a car) where the user must not be aware of the computational aspect. Our goal is to deeply analyze the architecture of the VISoc, a SoC developed by NeuriCam, exploiting its weaknesses, and improving it to achieve better performances.
Efficient data feeding for smart vision system
DANESE, GIOVANNI;GIACHERO, MAURO;LEPORATI, FRANCESCO;NAZZICARI, NELSON DAVIDE;
2005-01-01
Abstract
Single chip “smart” cameras have recently begun to appear and reach the market. These devices are complete System-on-Chip [SoC] capable of acquiring images at high speed and performing all the computation required to give users a set of high-level informations about the scene. The attractiveness of these devices comes from their low-cost, small-size and high-speed characteristics, making them particularly suitable for consumer applications (such as providing supplemental informations to drivers when installed on a car) where the user must not be aware of the computational aspect. Our goal is to deeply analyze the architecture of the VISoc, a SoC developed by NeuriCam, exploiting its weaknesses, and improving it to achieve better performances.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.