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

A general non-parametric technique is proposed for the analysis of a complex multimodal feature space and to delineate arbitrarily shaped clusters in it. The basic computational module of the technique is an old pattern recognition procedure: the mean shift. For discrete data, we prove the convergence of a recursive mean shift procedure to the nearest stationary point of the underlying density function and, thus, its utility in detecting the modes of the density. The relation of the mean shift procedure to the Nadaraya-Watson estimator from kernel regression and the robust M-estimators; of location is also established. Algorithms for two low-level vision tasks discontinuity-preserving smoothing and image segmentation - are described as applications. In these algorithms, the only user-set parameter is the resolution of the analysis, and either gray-level or color images are accepted as input. Extensive experimental results illustrate their excellent performance.

Mean shift: a robust approach toward feature space analysis

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

A survey on deep learning in medical image analysis

Computer vision applications have come to rely increasingly on superpixels in recent years, but it is not always clear what constitutes a good superpixel algorithm. In an effort to understand the benefits and drawbacks of existing methods, we empirically compare five state-of-the-art superpixel algorithms for their ability to adhere to image boundaries, speed, memory efficiency, and their impact on segmentation performance. We then introduce a new superpixel algorithm, simple linear iterative clustering (SLIC), which adapts a k-means clustering approach to efficiently generate superpixels. Despite its simplicity, SLIC adheres to boundaries as well as or better than previous methods. At the same time, it is faster and more memory efficient, improves segmentation performance, and is straightforward to extend to supervoxel generation.

/pdf/slic-superpixels-compared-to-state-of-the-art-superpixel-5d8sq04am9.pdf

SLIC Superpixels Compared to State-of-the-Art Superpixel Methods

The anatomy, function and hemodynamics of the aortic and mitral valves are known to be strongly interconnected. An integrated quantitative and visual assessment of the aortic-mitral coupling may have an impact on patient evaluation, planning and guidance of minimal invasive procedures. In this paper, we propose a novel model-driven method for functional and morphological characterization of the entire aortic-mitral apparatus. A holistic physiological model is hierarchically defined to represent the anatomy and motion of the two left heart valves. Robust learning-based algorithms are applied to estimate the patient-specific spatial-temporal parameters from four-dimensional TEE and CT data. The piecewise affine location of the valves is initially determined over the whole cardiac cycle using an incremental search performed in marginal spaces. Consequently, efficient spectrum detection in the trajectory space is applied to estimate the cyclic motion of the articulated model. Finally, the full personalized surface model of the aortic-mitral coupling is constructed using statistical shape models and local spatial-temporal refinement. Experiments performed on 65 4D TEE and 69 4D CT sequences demonstrated an average accuracy of 1.45mm and speed of 60 seconds for the proposed approach. Initial clinical validation on model-based and expert measurement showed the precision to be in the range of the inter-user variability. To the best of our knowledge this is the first time a complete model of the aortic-mitral coupling estimated from TEE and CT data is proposed.

/pdf/personalized-modeling-and-assessment-of-the-aortic-mitral-4da5jy16t3.pdf

Personalized Modeling and Assessment of the Aortic-Mitral Coupling from 4D TEE and CT

An artificial intelligence agent is machine trained and used to provide physically-based rendering settings. By using deep learning and/or other machine training, settings of multiple rendering parameters may be provided for consistent imaging even in physically-based rendering.

Content-based medical image rendering based on machine learning

A fetal parameter or anatomy is measured or detected from three-dimensional ultrasound data. An algorithm is machine-trained to detect fetal anatomy. Any machine training approach may be used. The machine-trained classifier is a joint classifier, such that one anatomy is detected using the ultrasound data and the detected location of another anatomy. The machine-trained classifier uses marginal space such that the location of anatomy is detected sequentially through translation, orientation and scale rather than detecting for all location parameters at once. The machine-trained classifier includes detectors for detecting from the ultrasound data at different resolutions, such as in a pyramid volume.

Automated Fetal Measurement From Three-Dimensional Ultrasound Data

Motion estimation for applications where appearance undergoes complex changes is challenging due to lack of an appropriate similarity function. In this paper, we propose to learn a discriminative similarity function based on an annotated database that exemplifies the appearance variations. We invoke the LogitBoost algorithm to selectively combine weak learners into one strong similarity function. The weak learners based on local rectangle features are constructed as nonparametric 2D piecewise constant functions, using the feature responses from both images, to strengthen the modeling power and accommodate fast evaluation. Because the negatives possess a location parameter measuring their closeness to the positives, we present a locationsensitive cascade training procedure, which bootstraps negatives for later stages of the cascade from the regions closer to the positives. This allows viewing a large number of negatives and steering the training process to yield lower training and test errors. In experiments of estimating the motion for the endocardial wall of the left ventricle in echocardiography, we compare the learned similarity function with conventional ones and obtain improved performances. We also contrast the proposed method with a learning-based detection algorithm to demonstrate the importance of temporal information in motion estimation. Finally, we insert the learned similarity function into a simple contour tracking algorithm and find that it reduces drifting.

/pdf/boostmotion-boosting-a-discriminative-similarity-function-4mr3jmlx38.pdf

BoostMotion: Boosting a Discriminative Similarity Function for Motion Estimation

Multiple candidates are optimized in medical device or feature tracking. Possible locations of medical devices or features for each of a plurality of different times are received. The possible locations of devices are modeled using a probability function. An iterative solution to obtain the maximum of the probability function determines the possible locations to be used as the locations of the medical devices or features for each time. Where two or more medical devices or features are provided with a geometric relationship, such as being connected by a detected guide wire, the probability function may account for the geometric relationship, such as a geodesic distance between the possible locations for the two medical devices.

Dorin Comaniciu

Papers

Personalized Modeling and Assessment of the Aortic-Mitral Coupling from 4D TEE and CT

Content-based medical image rendering based on machine learning

Automated Fetal Measurement From Three-Dimensional Ultrasound Data

BoostMotion: Boosting a Discriminative Similarity Function for Motion Estimation

Optimization of multiple candidates in medical device or feature tracking