The paper proposes the hardware implementation of the Gaussian Mixture Model (GMM) algorithm included in the OpenCV library. The OpenCV GMM algorithm is adapted to allow the FPGA implementation while providing a minimal impact on the quality of the processed videos. The circuit performs 30 frame per second (fps) background (Bg) identification on High Definition (HD) video sequences when implemented on commercial FPGA and outperforms previously proposed implementations. When implemented on Virtex5 lx50 FPGA using one level of pipeline, runs at 95.3 MHz, uses 5.3% of FPGA resources with a power dissipation of 1.47 mW/MHz.
FPGA implementation of OpenCV compatible background identification circuit / Genovese, Mariangela; Napoli, Ettore. - (2012), pp. 75-80. (Intervento presentato al convegno 3rd International Symposium on Computational Modelling of Objects Represented in Images: Fundamentals, Methods and Applications, CompIMAGE 2012 tenutosi a Rome (I) nel 5-7 Sept. 2012).
FPGA implementation of OpenCV compatible background identification circuit
GENOVESE, MARIANGELA;NAPOLI, ETTORE
2012
Abstract
The paper proposes the hardware implementation of the Gaussian Mixture Model (GMM) algorithm included in the OpenCV library. The OpenCV GMM algorithm is adapted to allow the FPGA implementation while providing a minimal impact on the quality of the processed videos. The circuit performs 30 frame per second (fps) background (Bg) identification on High Definition (HD) video sequences when implemented on commercial FPGA and outperforms previously proposed implementations. When implemented on Virtex5 lx50 FPGA using one level of pipeline, runs at 95.3 MHz, uses 5.3% of FPGA resources with a power dissipation of 1.47 mW/MHz.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
