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

Training a deep convolutional neural network (CNN) from scratch is difficult because it requires a large amount of labeled training data and a great deal of expertise to ensure proper convergence. A promising alternative is to fine-tune a CNN that has been pre-trained using, for instance, a large set of labeled natural images. However, the substantial differences between natural and medical images may advise against such knowledge transfer. In this paper, we seek to answer the following central question in the context of medical image analysis: Can the use of pre-trained deep CNNs with sufficient fine-tuning eliminate the need for training a deep CNN from scratch? To address this question, we considered four distinct medical imaging applications in three specialties (radiology, cardiology, and gastroenterology) involving classification, detection, and segmentation from three different imaging modalities, and investigated how the performance of deep CNNs trained from scratch compared with the pre-trained CNNs fine-tuned in a layer-wise manner. Our experiments consistently demonstrated that 1) the use of a pre-trained CNN with adequate fine-tuning outperformed or, in the worst case, performed as well as a CNN trained from scratch; 2) fine-tuned CNNs were more robust to the size of training sets than CNNs trained from scratch; 3) neither shallow tuning nor deep tuning was the optimal choice for a particular application; and 4) our layer-wise fine-tuning scheme could offer a practical way to reach the best performance for the application at hand based on the amount of available data.

Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?

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

IEEE Transactions on Pattern Analysis and Machine Intelligence

The WAVES investigation on the WIND spacecraft will provide comprehensive measurements of the radio and plasma wave phenomena which occur in Geospace. Analyses of these measurements, in coordination with the other onboard plasma, energetic particles, and field measurements will help us understand the kinetic processes that are important in the solar wind and in key boundary regions of the Geospace. These processes are then to be interpreted in conjunction with results from the other ISTP spacecraft in order to discern the measurements and parameters for mass, momentum, and energy flow throughout geospace. This investigation will also contribute to observations of radio waves emitted in regions where the solar wind is accelerated. The WAVES investigation comprises several innovations in this kind of instrumentation: among which the first use, to our knowledge, of neural networks in real-time on board a scientific spacecraft to analyze data and command observation modes, and the first use of a wavelet transform-like analysis in real time to perform a spectral analysis of a broad band signal.

/pdf/waves-the-radio-and-plasma-wave-investigation-on-the-wind-ngfb2pv1t7.pdf

Waves: The Radio and Plasma Wave Investigation on the Wind Spacecraft

A novel active contour model named variable background active contour model is proposed and applied for the detection of thyroid nodules in ultrasound images. The new model offers edge independency, no need for smoothing, ability for topological changes and it is more accurate when compared to the active contour without edges model. Improved accuracy is achieved by introducing as background a limited image subset which appropriately changes shape to reduce the effects of background inhomogeneity. We validated the proposed model on ultrasound images acquired from 24 patients and the results demonstrate an improvement in accuracy when compared to the active contour without edges model.

/pdf/a-variable-background-active-contour-model-for-automatic-3sd8n2333f.pdf

A variable background active contour model for automatic detection of thyroid nodules in ultrasound images

A few approaches have been presented in the literature towards the discrimination of texture in medical images. Medical experts proposed that the more valuable information for discriminating among normal and suspicious cancer regions in endoscopic images is the texture of the examined tissue. Texture can be encoded by a number of mathematical descriptors. Three well-known textural descriptors, as well as a new wavelet-based one are used in this paper for an accurate study and evaluation of the methodologies encountered. Experiments conducted include tests with various images from the Brodatz album, as well as interpretation of tissue regions in endoscopic image. In all cases the recognition task is supported by multilayer perceptron type neural network architectures.

Evaluation of textural feature extraction schemes for neural network-based interpretation of regions in medical images

Spot segmentation-the second essential stage of cDNA microarray image analysis-constitutes a challenging process. At present, several up-to-date spot-segmentation techniques or software programs-proposed in the literature-are often characterized as “automatic.” On the contrary, they are in effect not fully automatic since they require human intervention in order to specify mandatory input parameters or to correct their results. Human intervention, however, can inevitably modify the actual results of the cDNA microarray experiment and lead to erroneous biological conclusions. Therefore, the development of an automated spot-segmentation process becomes of exceptional interest. In this paper, an original and fully automatic approach to accurately segmenting the spots in a cDNA microarray image is presented. In order for the segmentation to be accomplished, each real spot of the cDNA microarray image is represented in a three-dimensional (3-D) space by a 3-D spot model. Each 3-D spot model is determined via an optimization problem, which is solved by using a genetic algorithm. The segmentation of real spots is conducted by drawing the contours of their 3-D spot models. The proposed method has been compared with various published and established techniques, using several synthetic and real cDNA microarray images that contain thousands of spots. The outcome has shown that the proposed method outperforms prevalent existing techniques. It is also noise resistant and yields excellent results even under adverse conditions such as the appearance of spots of various sizes and shapes.

/pdf/3-d-spot-modeling-for-automatic-segmentation-of-cdna-5c8wzsdpnm.pdf

3-D Spot Modeling for Automatic Segmentation of cDNA Microarray Images

In this work we compare two spatial and two wavelet-domain feature extraction methods that have been proposed in the recent literature for color-texture classification. The corresponding color-texture features, namely the Opponent-Color Local Binary Pattern distributions, the Chromaticity Moments, the Wavelet Correlation Signatures and the Color Wavelet Covariance features, are extracted in RGB, I1I2I3, HSV and CIE-Lab color spaces. The classification task is realized by Support Vector Machines. Experiments are performed on two standard datasets comprising of 54 and 68 textures from the Vistex and the Outex databases respectively. The results show that in most cases color enhances texture classification. Both spatial and wavelet features can lead to an accurate representation of color-textures. The appropriateness of a color-texture feature extraction method has to be determined by considering the trade-off between the accuracy and the feature space dimensionality needs, as these are imposed by a prospective application.

/pdf/a-comparative-study-of-color-texture-image-features-5g3oe24kgu.pdf

A comparative study of color- texture image features

This paper pertains to the recognition of texturalregions for color video analysis. The proposed schemeuses the covariance of 2nd-order statistics on the waveletdomain, between the different color channels of the videoframes. These features, named as Color WaveletCovariance (CWC), are used as color textural descriptors.A Support Vector Machine was chosen for theclassification of the CWC feature vectors. Experimentswere conducted using both animated Vistex texturemosaics and standard video clips. The estimated averageaccuracy ranged from 90% to 97%. The results show thatthe proposed methodology could efficiently be used invarious multimedia applications as a complete supervisedcolor texture recognition system.

/pdf/color-texture-recognition-in-video-sequences-using-wavelet-14748r734l.pdf

Dimitris Maroulis

Papers

A variable background active contour model for automatic detection of thyroid nodules in ultrasound images

Evaluation of textural feature extraction schemes for neural network-based interpretation of regions in medical images

3-D Spot Modeling for Automatic Segmentation of cDNA Microarray Images

A comparative study of color- texture image features

Color Texture Recognition in Video Sequences using Wavelet Covariance Features and Support Vector Machines