Course SEIA_205_1 · Year II · Autumn 2022-2023

Signal processing and pattern recognition.

Compulsory course in Advanced intelligent electronic systems, taught by Maria Loredana Oroian Boca.

This course page is from 2022-2023 and is archived.
See current courses for this faculty and study cycle

Overview

Lecturer
Maria Loredana Oroian Boca
Seminar tutor
Maria Loredana Oroian Boca
Type of course
Compulsory
Language of instruction
English
Erasmus language
English
Domain
Electronic engineering, telecommunication and information technologies - Masters
Field of study
Advanced intelligent electronic systems
Form of education
Full-time
Form of instruction
Lecture
Credit awarded by
Grade
Teaching methods
lectures, examples, exercises
Entry requirements
Artificial Intelligence, machines and shape recognition

Aims

Acquire knowledge about analysis and interpretation of data and information

Development of hardware and software applications for the field of intelligent systems and industrial informatics by choosing the optimal solution, designing a functional and integrated testing plan, interpreting the results, comparing them with the expected ones and developing the correction method

Processing of complex signals (voice, data, text, images), with encryption, compression

Modeling, implementing, testing, using and maintaining advanced electronic systems.

The applications aim to familiarize students with the general techniques of form recognition. Students must design (at least in Matlab) applications for the recognition of the following fields:  medical diagnosis;  automotive (pedestrian recognition, traffic signs, etc.);  biometrics;

Course contents

1. Introduction to the theory of machine tools. The learning process. -Learning in neural networks - Assessing hypotheses and Bayesian learning - Instance-based learning. - Machines with vector support 2.Neural networks. Feed-forward neural networks (with supervised learning). -Multilayer Perceptron (MLP); - Networks with basic radial functions (FBR);Recurrent neural networks (with supervised learning). - The Hopfield Network; - Two-way associative memory (BAM); Competitive neural networks (with unsupervised learning). - Self-Organizing Maps (SOM); - Networks based on adaptive resonance theory (ART); 4. Fuzzy and Neuro-Fuzzy systems - Introduction to nuanced logic ("fuzzy"). "Fuzzy" relationships. The metric "fuzzy". "Fuzzy" implications. Approximate reasoning. Classification systems with "fuzzy" rules. - Neuro-fuzzy networks. Integration of "fuzzy" logic and neural networks. Fuzzy neurons. 5. Genetic algorithms. - The stages of a genetic algorithm: selection, crossing, mutation. - Neural networks with genetic algorithms. - Applications in shape recognition 6. Two-dimensional signal processing elements (images) 7. Recognition of shapes in images - Statistical methods. Bayes classifier, - Classification based on prototype, - The kNN classifier, 8. Applications. - Classification of medical signals and images. - Classification of planar objects. - Biometric technology, iris recognition, face identification. - Artificial view mobile robots

Learning outcomes

Development of programs in a general and / or specific programming language, starting from the specification of the requirements and until the execution, debugging and interpretation of the results.

Assessment

projects

Recommended reading

Machine Learning
Mitchell, T
The McGraw-Hill Companies, 1997
IMAGE PROCESSING AND PATTERN RECOGNITION Fundamentals and Techniques
FRANK Y. SHIH
Published by John Wiley & Sons, Inc., Hoboken,, New Jersey., 2010
Understanding Digital Processing
Richard G Lyons
Pearson Education, Inc., 2011
Digital Signals Processing
Steven W Smith
California Technical Publishing, 1999