Difference between revisions of "Applications of Pattern Recognition Old Kiwi" - Rhea

Revision as of 11:16, 11 April 2008

This page can be used to discuss the applications of pattern recognition in our daily research! This would provide us an intuitive understanding of course topics. Please discuss "applied" pattern recognition here. Instead of just mentioning the field, please explain in detail how a specific tool of pattern recognition can be used in research.

• Case-based Reasoning_Old Kiwi
• Wireless Communications_Old Kiwi
• Image Processing_Old Kiwi
• Implementation Issues_Old Kiwi
• Video Classification - State of the Art_Old Kiwi

From yamini.nimmagadda.1 Sun Feb 3 15:44:44 -0500 2008 From: yamini.nimmagadda.1 Date: Sun, 03 Feb 2008 15:44:44 -0500 Subject: Maximum Likelihood Estimate Message-ID: <20080203154444-0500@https://engineering.purdue.edu>

In Wireless Communications:

• If the input sequence messages are equally likely, Convolutional decoders like Viterbi minimizes the probability of error using the maximum likelihood estimate between the output sequence and all the possible input sequences.

In Image Processing:

• Maximum likelihood estimates can be used in image matching (edge template matching and gray-level image matching). This can be applied to stereo matching and feature tracking. More about this topic can be found here ... [1]
• Maximum likelihood can also be used in image reconstruction or restoration. Surprisingly, I found the usage of this in compression artifact removal also. See this paper [2]

In Signal Processing:

• automatic speech recognition
• face recognition

In Face Reconition:

• Fisher Linear Discriminant (FLD) is widely used in face recognition. Here is a paper for reference: [3]. Also variants of FLD are used for face recognition such as DiaFLD [4]. It has been observed that FLD works better than Principal Component Analysis in classifying the facial features.

Face detection vs Face recognition_Old Kiwi

In Image Segmentation:

• Image Segmentation is performed by conventional graphical methods, but many a times, some pixels not belonging to the same object are classified into the same segment. Also, in images where a wide background is separated by a thin boundary line, image segmentation can be performed by obtaining features from FLD. I experimented this personally and found that the results are better than the conventional methods. This paper gives a starting point in doing this [5].

In Artifical Intelligence:

• Case-based Reasoning_Old Kiwi

The applications of support vector machines in various fields can be found here. [6]

In Computer Security

Anomaly Detection for Computer Security (taken verbatim from http://www.cs.unm.edu/~terran/research/anomaly_detection_for_computer_security)

".. A number of critical problems in computer security can be viewed as distinguishing some "normal" circumstance from "anomalous" or "abnormal" circumstances. For example, we can think of computer viruses as being (syntactic and begavioral) abnormal modifications to normal programs. Similarly, network intrusion detection is also an attempt to discern unusual or abnormal patterns in network traffic. Superficially, this is a standard binary concept learning problem from supervised learning. In practice, however, it's usually infeasible to treat the problem directly this way. Typically, we don't have a thorough sample of examples of abnormal/hostile data, either because the data itself is hard to come by (many sources don't preserve or won't release records of their own vulnerabilities) or because novel attacks are constantly being introduced. Furthermore, defenses based on any fixed assumption of the distribution of attacks would be vulnerable to attacks designed specifically to subvert that assumption. (Virus authors, for example, appear to test their new strains against current commercial antivirus programs in order to develop undetectable strains.) Thus, it is often advantageous to conceive of the anomaly detection problem as the task of developing a strong model of normal behaviors and detecting abnormalities as deviations from that model. This offers the dual benefits of adaptivity to individual systems/users/sites and of (in principle) being less vulnerable to novel attacks.

A bit more formally, the anomaly detection problem can be framed as a distribution estimation problem for a single class of data (normal behavior) coupled with a threshold selection procedure to define the negative pattern (anomaly) space. The challenge lies in developing sufficiently descriptive models of normal behavior that still allow discrimination of abnormalities .."