Title :
Fast Argument Quantization [SP Tips&Tricks]
Author :
Nilsson, John-Olof ; Handel, Peter
Author_Institution :
Dept. of Signal Process., KTH R. Inst. of Technol., Stockholm, Sweden
Abstract :
A quantized low-resolution argument of a complex number or two-dimensional vector is required in many digital signal processing algorithms. Examples include APSK code demodulation for which it may be used to evaluate the Voronoi diagram; low-level processing for many computer vision methods that exploit histograms of gradient sample arguments, e.g., SIFT and HOG; and phase tracking/frequency estimation for which it may be used as a low-cost phase approximation. Often, such quantized arguments will have to be computed many times and under real-time constraints. Therefore, efficient solutions to these calculations are of interest.
Keywords :
quantisation (signal); signal processing; APSK code demodulation; Voronoi diagram; computer vision methods; digital signal processing algorithms; fast argument quantization; frequency estimation; gradient sample arguments; low cost phase approximation; low level processing; phase tracking; quantized low resolution argument; two dimensional vector; Approximation methods; Mobile communication; Quantization (signal); Real-time systems; Streaming media; Ultrasonic imaging;
Journal_Title :
Signal Processing Magazine, IEEE
DOI :
10.1109/MSP.2013.2276531