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

Biological conclusions reached during microarray experiments can be greatly affected by human intervention, which is currently required in microarray image analysis. Therefore, accurate and automatic analysis of cDNA microarray images becomes crucial. In this paper, an automatic approach to microarray image analysis is presented. The proposed approach is based on the concept of evolution in order to process the microarray images. Conducted experiments in a set of real microarray images confirm the effectiveness of the proposed approach.

Microarray image analysis based on an evolutionary approach

A miniaturized laser scanning endoscope is presented which makes use of three lasers to illuminate a sample with a red, a green and a blue wavelength simultaneously. Scattered light from the sample is descanned and chromatically separated into the three channels for detection and postprocessing to compose a color image. The scanning subsystem consists of two micro-electro-mechanical mirrors suitable for mass production. The endoscope head can be assembled fast and at low cost. A resolution of the order of 16 lines per mm is achieved for a working distance common in endoscopy. Considerations of the system design include the operation of the mico mirrors, the filtering of reflected light by using polarization effects and a strategy to cope with color metamery. An expert system based on a neural network was found able to analyze endoscopic images to identify suspicious lesions.

A laser-scanning endoscope based on polysilicon micromachined mirrors with enhanced attributes

In this chapter, a variant of the Local Binary Patterns method that extends the ability of the method to cope with noisy texture by utilizing fuzzy modelling techniques is presented. The presented generalised Fuzzy Binary Patterns model is applied to the classic Local Binary Patterns method as well as to the Local Binary Patterns with Contrast measure method, resulting to the respective fuzzy logic based methods. Supervised classification experiments were conducted on a wide range of natural and medical texture images, degraded by different types and intensities of additive noise, in order to evaluate the efficiency of the Fuzzy Local Binary Patterns method and its fusion with other proposed methods. Fuzzy Binary Patterns based methods outperform the respective methods based on the classic Binary Patterns model for all types of images and noise, indicating the efficiency of fuzzy modelling in coping with the uncertainty introduced to texture due to noise.

FLBP: Fuzzy Local Binary Patterns

This work addresses the segmentation of two-dimensional polyacrylamide gel electrophoresis images containing overlapping protein spots. A novel segmentation approach is proposed, which is capable of detecting spot boundaries within the region of overlap. The proposed approach is based on the observation that the spot boundaries in the overlap region are associated with local intensity minima. The experimental evaluation of the proposed approach demonstrates that it is capable of identifying multiple overlapping spots, as opposed to state of the art segmentation approaches. Moreover, it results in more accurate spot delineations, when compared to Progenesis SameSpots.

/pdf/segmentation-of-two-dimensional-gel-electrophoresis-images-22gxjnhg1p.pdf

Segmentation of two-dimensional gel electrophoresis images containing overlapping spots

This work introduces a novel framework for unsupervised parameterization of region-based active contour regularization and data fidelity terms, which is applied for medical image segmentation. The work aims to relieve MDs from the laborious, time-consuming task of empirical parameterization and bolster the objectivity of the segmentation results. The proposed framework is inspired by an observed isomorphism between the eigenvalues of structure tensors and active contour parameters. Both may act as descriptors of the orientation coherence in regions containing edges. The experimental results demonstrate that the proposed framework maintains a high segmentation quality without the need of trial-and-error parameter adjustment.

/pdf/self-parameterized-active-contours-based-on-regional-edge-8j8bs8vl47.pdf

Dimitris Maroulis

Papers

Microarray image analysis based on an evolutionary approach

A laser-scanning endoscope based on polysilicon micromachined mirrors with enhanced attributes

FLBP: Fuzzy Local Binary Patterns

Segmentation of two-dimensional gel electrophoresis images containing overlapping spots

Self-parameterized active contours based on regional edge structure for medical image segmentation