Results 1 
4 of
4
Adaptive Scalar Quantization without Side Information
 IEEE Trans. Image Proc
, 1997
"... In this paper, we introduce a novel technique for adaptive scalar quantization. Adaptivity is useful in applications, including image compression, where the statistics of the source are either not known a priori or will change over time. Our algorithm uses previously quantized samples to estimate th ..."
Abstract

Cited by 18 (4 self)
 Add to MetaCart
In this paper, we introduce a novel technique for adaptive scalar quantization. Adaptivity is useful in applications, including image compression, where the statistics of the source are either not known a priori or will change over time. Our algorithm uses previously quantized samples to estimate the distribution of the source, and does not require that side information be sent in order to adapt to changing source statistics. Our quantization scheme is thus backward adaptive. We propose that an adaptive quantizer can be separated into two building blocks, namely, model estimation and quantizer design. The model estimation produces an estimate of the changing source probability density function, which is then used to redesign the quantizer using standard techniques. We introduce nonparametric estimation techniques that only assume smoothness of the input distribution. We discuss the various sources of error in our estimation and argue that, for a wide class of sources with a smooth probability density function (pdf), we provide a good approximation to a "universal" quantizer, with the approximation becoming better as the rate increases. We study the performance of our scheme and show how the loss due to adaptivity is minimal in typical scenarios. In particular, we provide examples and show how our technique can achieve signalto noise ratios (SNR's) within 0.05 dB of the optimal LloydMax quantizer (LMQ) for a memoryless source, while achieving over 1.5 dB gain over a fixed quantizer for a bimodal source.
Sensory Flow Segmentation using a Resource Allocating Vector Quantizer
, 2000
"... . We present a very simple unsupervised vector quantizer which extracts higher order concepts from time series generated from sensors on a mobile robot as it moves through an environment. The vector quantizer is constructive, i.e. it adds new model vectors, each one encoding a separate higher or ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
. We present a very simple unsupervised vector quantizer which extracts higher order concepts from time series generated from sensors on a mobile robot as it moves through an environment. The vector quantizer is constructive, i.e. it adds new model vectors, each one encoding a separate higher order concept, to account for any novel situation the robot encounters. The number of higher order concepts is determined dynamically, depending on the complexity of the sensed environment, without the need of any user intervention. We show how the vector quantizer elegantly handles many of the problems faced by an existing architecture by Nol and Tani, and note some directions for future work. 1 Introduction As a mobile robot moves through an environment, it receives a sequence of inputs through its sensory equipment, this sequence of inputs is called the `sensory ow'. The sensory ow can easily be in the order of thousands, or even millions, of discrete samples. Finding relations an...
Extraction and Inversion of Abstract Sensory Flow Representations
, 2000
"... We present a technique for analysing extracted abstract sensory ow representations. Mobile robot systems generally have a very limited memory storage capacity, and can thus only store information about a very short period of time. However, instead of storing a very detailed picture of the past ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
We present a technique for analysing extracted abstract sensory ow representations. Mobile robot systems generally have a very limited memory storage capacity, and can thus only store information about a very short period of time. However, instead of storing a very detailed picture of the past, a more abstract representation can be extracted and stored, providing a much longer, but less detailed, picture of the past to the control system. In our technique, the components of such an abstract representation are automatically extracted; they correspond to distinct and stable inputs which arrive at the sensory systems. To a distal observer, the extracted components actually correspond to concepts such as corridors, walls and corners. However, the analysis of the extracted components has previously required that the experimenter constantly watches as the robot does something and notes which concepts get activated. This is cumbersome, and such an inspection may even be i...
Self Organizing Map for Adaptive Nonstationary Clustering: some experimental results on Color Quantization of image sequences
"... : In this paper we consider the application of the Self Organizing Map to the adaptive computation of cluster representatives (codevectors) over nonstationary data. The paradigm of Nonstationary Clustering is represented by the problem of Color Quantization of image sequences. Experimental results ..."
Abstract
 Add to MetaCart
: In this paper we consider the application of the Self Organizing Map to the adaptive computation of cluster representatives (codevectors) over nonstationary data. The paradigm of Nonstationary Clustering is represented by the problem of Color Quantization of image sequences. Experimental results on the Color Quantization of an image sequence show the extreme robustness of the SOM as an adaptive clustering algorithm. 0 Introduction Cluster analysis and Vector Quantization have applications in signal processing, pattern recognition, machine learning and data analysis [1,2,3,4,5,6]. A vast number of approaches have been proposed to solve these problems, among them Competitive Neural Networks have been proposed as a kind of adaptive partitional methods [7,8,9]. Conventional formulations of Clustering and Vector Quantization assume that the underlying stochastic process is stationary and that a given set of sample vectors properly characterizes this process. Nonstationary processes ar...