Title :
Improving the performance of MPEG compatible encoders using on line retrainable neural networks
Author :
Kollias, Stefanos ; Doulamis, Nikolaos ; Doulamis, Anastasios
Author_Institution :
Electr. & Comput. Eng., Nat. Tech. Univ. of Athens, Greece
Abstract :
On line retraining of neural network is introduced for extracting foreground/background objects in video sequences. The scheme is applied together with a modification of the rate control of MPEG-1 algorithm. The proposed method is compatible to MPEG-1/2 standard but also can be used as a pre-coding stage for the forthcoming MPEG-4 algorithm. Simulation studies have shown an improvement of about 1.5 dB on average as far the PSNR is concerned compared with the conventional MPEG-1 encoder
Keywords :
code standards; feature extraction; image sequences; learning (artificial intelligence); neural nets; real-time systems; telecommunication standards; video coding; MPEG compatible encoders; MPEG-1 algorithm; MPEG-1 encoder; MPEG-1/2 standard; MPEG-4 algorithm; PSNR; foreground/background objects extraction; online retrainable neural networks; performance; precoding stage; rate control; simulation; video sequences; Computer networks; Decoding; Humans; Image coding; Image quality; Image segmentation; MPEG 4 Standard; Neural networks; Transform coding; Videoconference;
Conference_Titel :
Image Processing, 1997. Proceedings., International Conference on
Conference_Location :
Santa Barbara, CA
Print_ISBN :
0-8186-8183-7
DOI :
10.1109/ICIP.1997.632146
