Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

MUCKE aims to mine a large volume of images, to structure them conceptually and to use this conceptual structuring in order to improve large-scale image retrieval. The last decade witnessed important progress concerning low-level image representations. However, there are a number problems which need to be solved in order to unleash the full potential of image mining in applications. The central problem with low-level representations is the mismatch between them and the human interpretation of image content. This problem can be instantiated, for instance, by the incapability of existing descriptors to capture spatial relationships between the concepts represented or by their incapability to convey an explanation of why two images are similar in a content-based image retrieval framework. We start by assessing existing local descriptors for image classification and by proposing to use co-occurrence matrices to better capture spatial relationships in images. The main focus in MUCKE is on cleaning large scale Web image corpora and on proposing image representations which are closer to the human interpretation of images. Consequently, we introduce methods which tackle these two problems and compare results to state of the art methods. Note: some aspects of this deliverable are withheld at this time as they are pending review. Please contact the authors for a preview.

http://ieeexplore.ieee.org/iel2/4524/12820/00592460.pdf

Image Processing

A scheme is developed for classifying the types of motion perceived by a humanlike robot. It is assumed that the robot receives visual images of the scene using a perspective system model. Equations, theorems, concepts, clues, etc., relating the objects, their positions, and their motion to their images on the focal plane are presented. >

http://ieeexplore.ieee.org/iel2/815/3148/00100651.pdf

Robot vision

Computer and Robot Vision Vol. 1, by R.M. Haralick and Linda G. Shapiro, Addison-Wesley, 1992, ISBN 0-201-10887-1.

Computer and Robot Vision

This chapter introduces the subject of statistical pattern recognition (SPR). It starts by considering how features are defined and emphasizes that the nearest neighbor algorithm achieves error rates comparable with those of an ideal Bayes’ classifier. The concepts of an optimal number of features, representativeness of the training data, and the need to avoid overfitting to the training data are stressed. The chapter shows that methods such as the support vector machine and artificial neural networks are subject to these same training limitations, although each has its advantages. For neural networks, the multilayer perceptron architecture and back-propagation algorithm are described. The chapter distinguishes between supervised and unsupervised learning, demonstrating the advantages of the latter and showing how methods such as clustering and principal components analysis fit into the SPR framework. The chapter also defines the receiver operating characteristic, which allows an optimum balance between false positives and false negatives to be achieved.

Statistical Pattern Recognition

Several color object recognition methods that are based on image retrieval algorithms attempt to discount changes of illumination in order to increase performance when test image illumination conditions differ from those that obtained when the image database was created. Here we extend the seminal method of Swain and Ballard to discount changing illumination. The new method is based on the first stage of the simplest color indexing method, which uses angular invariants between color image and edge image channels. That method first normalizes image channels, and then effectively discards much of the remaining information. Here we adopt the color-normalization stage as an adequate color constancy step. Further, we replace 3D color histograms by 2D chromaticity histograms. Treating these as images, we implement the method in a compressed histogram-image domain using a combination of wavelet compression and Discrete Cosine Transform (DCT) to fully exploit the technique of low-pass filtering for efficiency. Results are very encouraging, with substantially better performance than other methods tested. The method is also fast, in that the indexing process is entirely carried out in the compressed domain and uses a feature vector of only 36 or 72 values.

Illumination-invariant color object recognition via compressed chromaticity histograms of color-channel-normalized images

Illumination–invariant image retrieval and video segmentation

Edit distance was originally developed by Levenstein several decades ago to measure the distance between two strings. It was found that this distance can be computed by an elegant dynamic programming procedure. The edit distance has played important roles in a wide array of applications due to its representational efficacy and computational efficiency. To effect a more reasonable distance measure, the normalized edit distance was proposed. Many algorithms and studies have been dedicated along this line with impressive performances in recent years. There is, however, a fundamental problem with the original definition of edit distance that has remained elusive: its context-free nature. In determining the possible actions, i.e., insertion, deletion, and substitution, no systematic consideration was given to the local behaviors of the string/pattern in question that indeed encompass great amount of useful information regarding its content. In this paper, inspired by the success of the Markov Random Field theory, a new edit distance called Markov edit distance (MED) within the dynamic programming framework is proposed to take full advantage of the local statistical dependencies in the pattern in order to arrive at enhanced matching performance. Within this framework, two specialized distance measures are developed: The reshuffling MED to handle cases where a subpattern in the target pattern is the reshuffles of that in the source pattern, and the coherence MED which is able to incur local content based substitution, insertion, and deletion. The applications based on these two MEDs in string matching are then explored, whereof encouraging empirical results have been observed.

Markov edit distance

On illumination invariance in color object recognition

Purpose To compare the image quality of amplitude-binned 4-dimensional magnetic resonance imaging (4DMRI) reconstructed using 2 concurrent respiratory (navigator and bellows) waveforms. Methods and Materials A prospective, respiratory-correlated 4DMRI scanning program was used to acquire T2-weighted single-breath 4DMRI images with internal navigator and external bellows. After a 10-second training waveform of a surrogate signal, 2-dimensional MRI acquisition was triggered at a level (bin) and anatomic location (slice) until the bin-slice table was completed for 4DMRI reconstruction. The bellows signal was always collected, even when the navigator trigger was used, to retrospectively reconstruct a bellows-rebinned 4DMRI. Ten volunteers participated in this institutional review board–approved 4DMRI study. Four scans were acquired for each subject, including coronal and sagittal scans triggered by either navigator or bellows, and 6 4DMRI images (navigator-triggered, bellows-rebinned, and bellows-triggered) were reconstructed. The simultaneously acquired waveforms and resulting 4DMRI quality were compared using signal correlation, bin/phase shift, and binning motion artifacts. The consecutive bellows-triggered 4DMRI scan was used for indirect comparison. Results Correlation coefficients between the navigator and bellows signals were found to be patient-specific and inhalation-/exhalation-dependent, ranging from 0.1 to 0.9 because of breathing irregularities (>50% scans) and commonly observed bin/phase shifts (−1.1 ± 0.6 bin) in both 1-dimensional waveforms and diaphragm motion extracted from 4D images. Navigator-triggered 4DMRI contained many fewer binning motion artifacts at the diaphragm than did the bellows-rebinned and bellows-triggered 4DMRI scans. Coronal scans were faster than sagittal scans because of the fewer slices and higher achievable acceleration factors. Conclusions Navigator-triggered 4DMRI contains substantially fewer binning motion artifacts than bellows-rebinned and bellows-triggered 4DMRI, primarily owing to the deviation of the external from the internal surrogate. The present study compared 2 concurrent surrogates during the same 4DMRI scan and their resulting 4DMRI quality. The navigator-triggered 4DMRI scanning protocol should be preferred to the bellows-based, especially for coronal scans, for clinical respiratory motion simulation.

Direct Comparison of Respiration-Correlated Four-Dimensional Magnetic Resonance Imaging Reconstructed Using Concurrent Internal Navigator and External Bellows.

Jie Wei

Papers

Illumination-invariant color object recognition via compressed chromaticity histograms of color-channel-normalized images

Illumination–invariant image retrieval and video segmentation

Markov edit distance

On illumination invariance in color object recognition

Direct Comparison of Respiration-Correlated Four-Dimensional Magnetic Resonance Imaging Reconstructed Using Concurrent Internal Navigator and External Bellows.