TY - JOUR
T1 - Design and analysis of binary scalar quantizer of laplacian source with applications
AU - Peric, Zoran
AU - Denic, Bojan
AU - Savic, Milan
AU - Despotovic, Vladimir
N1 - Funding Information:
Funding: This work has been supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia and by the Science Found of the Republic of Serbia (Grant No. 6527104, AI-Com-in-AI).
Publisher Copyright:
© 2020 by the authors. Licensee MDPI, Basel, Switzerland.
PY - 2020/11
Y1 - 2020/11
N2 - A compression method based on non-uniform binary scalar quantization, designed for the memoryless Laplacian source with zero-mean and unit variance, is analyzed in this paper. Two quantizer design approaches are presented that investigate the effect of clipping with the aim of reducing the quantization noise, where the minimal mean-squared error distortion is used to determine the optimal clipping factor. A detailed comparison of both models is provided, and the performance evaluation in a wide dynamic range of input data variances is also performed. The observed binary scalar quantization models are applied in standard signal processing tasks, such as speech and image quantization, but also to quantization of neural network parameters. The motivation behind the binary quantization of neural network weights is the model compression by a factor of 32, which is crucial for implementation in mobile or embedded devices with limited memory and processing power. The experimental results follow well the theoretical models, confirming their applicability in real-world applications.
AB - A compression method based on non-uniform binary scalar quantization, designed for the memoryless Laplacian source with zero-mean and unit variance, is analyzed in this paper. Two quantizer design approaches are presented that investigate the effect of clipping with the aim of reducing the quantization noise, where the minimal mean-squared error distortion is used to determine the optimal clipping factor. A detailed comparison of both models is provided, and the performance evaluation in a wide dynamic range of input data variances is also performed. The observed binary scalar quantization models are applied in standard signal processing tasks, such as speech and image quantization, but also to quantization of neural network parameters. The motivation behind the binary quantization of neural network weights is the model compression by a factor of 32, which is crucial for implementation in mobile or embedded devices with limited memory and processing power. The experimental results follow well the theoretical models, confirming their applicability in real-world applications.
KW - Clipping factor
KW - Delta modulation
KW - Laplacian distribution
KW - Neural network
KW - Pulse code modulation
KW - Quantization
KW - Signal to noise ratio
UR - http://www.scopus.com/inward/record.url?scp=85094636625&partnerID=8YFLogxK
U2 - 10.3390/info11110501
DO - 10.3390/info11110501
M3 - Article
AN - SCOPUS:85094636625
SN - 2078-2489
VL - 11
SP - 1
EP - 18
JO - Information (Switzerland)
JF - Information (Switzerland)
IS - 11
M1 - 501
ER -