Cellular Associative Neural Networks (CANN)

نویسندگان

  • Grant Brewer
  • Stefan Klinger
چکیده

We propose to present a novel syntactic pattern matching technique [1] that combines the correlative learning and generalisation properties of associative memories, with the parallel and distributed operation of cellular automata [2]. The tool is used for recognizing two dimensional objects in images. Each section of the object to be studied is represented by a cell, initially containing a low level symbolic representation of that section. Each cell is then iteratively updated, based on its previous state and that of its neighbours, giving a gradually higher level representation of the area, until an object level definition can be obtained. The update rules are stored in Advanced Uncertain Reasoning Architecture (AURA) [3] associative memories, allowing efficient relaxation in situations where exact matches are unavailable, such as in the presence of a noisy input. In addition to an overview of the method, current applications of this method will be highlighted, including graph matching and noisy image recognition. Some of problems and challenges faced by the technique will also be discussed.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A cross-associative neural network for SVD of non-squared data matrix in signal processing

This paper proposes a cross-associative neural network (CANN) for singular value decomposition (SVD) of a non-squared data matrix in signal processing, in order to improve the convergence speed and avoid the potential instability of the deterministic networks associated with the cross-correlation neural-network models. We study the global asymptotic stability of the network for tracking all the...

متن کامل

A Self-Reconstructing Algorithm for Single and Multiple-Sensor Fault Isolation Based on Auto-Associative Neural Networks

Recently different approaches have been developed in the field of sensor fault diagnostics based on Auto-Associative Neural Network (AANN). In this paper we present a novel algorithm called Self reconstructing Auto-Associative Neural Network (S-AANN) which is able to detect and isolate single faulty sensor via reconstruction. We have also extended the algorithm to be applicable in multiple faul...

متن کامل

Evolutionary Techniques Based Combined Artificial Neural Networks for Peak Load Forecasting

This paper presents a new approach using Combined Artificial Neural Network (CANN) module for daily peak load forecasting. Five different computational techniques –Constrained method, Unconstrained method, Evolutionary Programming (EP), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA) – have been used to identify the CANN module for peak load forecasting. In this paper, a set of ne...

متن کامل

Multi-packet regions in stabilized continuous attractor networks

Continuous attractor neural networks are recurrent networks with center-surround interaction profiles which are common ingredients in many neuroscientific models. The basic CANN model is often augmented with mechanisms reflecting activitydependent cellular nonlinearities. In this paper, we study the balance between global competition and the stabilizing effects of cellular nonlinearities, and d...

متن کامل

Cloning Template Design of Cellular Neural Networks for Associative Memories

This brief presents a synthesis procedure (design algorithm) for cellular neural networks with space-invariant cloning template with applications to associative memories. The design algorithm makes it possible to determine in a systematic manner cloning templates for cellular neural networks with or without symmetry constraints on the interconnection weights. Two specific examples are included ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2005