Tutorial 1

Advanced classification techniques for image and text analysis based on clustering

Dr. Parag Kulkarni, GM-Research, Capsilon India, Kalyani Nagar, Pune, 411 014 INDIA E-mail: paragakulkarni@yahoo.com

Dr. Parag Kulkarni (paragakulkarni@yahoo.com) is Ph.D. form IIT, Kharagpur (www.iitkgp.ernet.in). He is working in IT industry for more than 13 years. He is on research panel and Ph.D. guide for University of Pune, BITS and Symbiosis deemed University. He has conducted 5 tutorials at various international conferences and was a keynote speaker for three international conferences.
He has also worked as a referee for International Journal for Parallel and Distributed Computing, IASTED conferences. He is also member of IASTED technical committee of Parallel and Distributed Computing. Presently he is General Manager-Technical at Capsilon India, Pune. He is also Honorary Professor at AISSM Engineering College, Pune and on board of studies for a couple of Institutes. He has worked as Senior Manager (R&D) at Scientific Applications Center, Siemens information systems Ltd., Pune. His areas of interest include image processing, security systems, decision systems, expert systems, classification techniques, load balancing and distributed computing.

Scope

In all learning techniques and decision methods we need to classify various objects, behaviors and patterns. In this tutorial we will discuss advanced classification algorithms in detail. We will talk about various pattern and behavior based methods used for text and image classification. The tutorial will also cover usage of various methods like SVM, intelligent feature extraction; feature based clustering for these applications. In this tutorial we will cover various difficulties in classification and learning methodologies,which can handle nonlinear behavior very effectively. The tutorial will also present industrial applications of classification and case studies for classifying images and textual data. It will also throw some light on how these techniques can be extended for other applications in decision engineering.

Objectives:

This tutorial is formulated with following objectives:

To highlight importance of clustering and classification in decision making
Analyze traditional as well as novel methods of classification.
To provide a common thread that works across various tools and leads to decision engineering
Highlight various key Applications of classification

Introduction

When I was working with decision engineering, decision support systems for almost 13 years in various industries, I realized that these all things could be put together and can be shared with various audiences. That gave me inspiration to present tutorial and I started sharing experiences by means of tutorials and learned a lot through this exercise.

Classification is complex process. Classification is simple as long as it is linear. But most of the problems are non-linear and there is lot of noise. While applying the techniques we need to work on various parameters and it also requires tuning and tweaking. Here in this tutorial we are going to present step-by-step approach for classification and which ultimately leads to decision-making or can be part of decision support systems. We will also talk about clustering, outlier removal and other real life scenario without which the discussion may not be completed.

Duration: 4.00 HR

Program:

1. Introduction

What is classification?
Various classification methods.
Classification and Decision Support systems

2. Intelligent feature extraction for classification

Behavioral patterns
Behavioral patterns of images and objects
Clustering based classification
Tracking changes in behavior and learning.

3. Various methods for classification

Support vector machines
SVM for image classification
SVM for text classification
Nonlinear behavior pattern and SVM

4. Case studies

Industrial applications
Face authentication
Image based change detection

5. Applications

Industrial applications
Decisions based on classifications
Learning based on classification

6. Concluding remarks.

Target Audience

Professionals working in area of decision systems, classification, image processing
Researchers working in area of decision systems, classifications, image processing, statistical techniques
Students with some background in decision system and want to pursue career in these area

Background

Some background in statistics
Exposure to AI and neural networks
Some knowledge about basic subjects of computers, neural networks and mathematical foundation of clustering is desirable but not mandetory