학술논문

A new object-oriented segmentation algorithm based on CNNs - part II: performance evaluation
Document Type
Conference
Source
2005 9th International Workshop on Cellular Neural Networks and Their Applications Cellular Neural Networks and their Applications Cellular Neural Networks and Their Applications, 2005 9th International Workshop on. :150-153 2005
Subject
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Cellular neural networks
Image segmentation
Video sequences
Object oriented modeling
Hardware
Gray-scale
Filtering
Image processing
MPEG 4 Standard
Computer networks
Language
ISSN
2165-0144
2165-0152
Abstract
By using the CNN paradigm, this paper illustrates a new object-oriented segmentation algorithm that takes into account the hardware characteristics imposed by the CNNUM. In particular, this paper describes every block of the algorithm except the edge extraction one, which is described in the companion paper (Grasi et al., 2005). Additionally, by considering different video sequences, this paper illustrates some performance evaluations, showing that the approach (based on a rigorous model of the image contours) provides more accurate segmented objects than the ones obtained by other CNN-based techniques.