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

http://library.utia.cas.cz/prace/20030125.pdf

Image registration methods: a survey

Humans perceive the three-dimensional structure of the world with apparent ease. However, despite all of the recent advances in computer vision research, the dream of having a computer interpret an image at the same level as a two-year old remains elusive. Why is computer vision such a challenging problem and what is the current state of the art? Computer Vision: Algorithms and Applications explores the variety of techniques commonly used to analyze and interpret images. It also describes challenging real-world applications where vision is being successfully used, both for specialized applications such as medical imaging, and for fun, consumer-level tasks such as image editing and stitching, which students can apply to their own personal photos and videos. More than just a source of recipes, this exceptionally authoritative and comprehensive textbook/reference also takes a scientific approach to basic vision problems, formulating physical models of the imaging process before inverting them to produce descriptions of a scene. These problems are also analyzed using statistical models and solved using rigorous engineering techniques Topics and features: structured to support active curricula and project-oriented courses, with tips in the Introduction for using the book in a variety of customized courses; presents exercises at the end of each chapter with a heavy emphasis on testing algorithms and containing numerous suggestions for small mid-term projects; provides additional material and more detailed mathematical topics in the Appendices, which cover linear algebra, numerical techniques, and Bayesian estimation theory; suggests additional reading at the end of each chapter, including the latest research in each sub-field, in addition to a full Bibliography at the end of the book; supplies supplementary course material for students at the associated website, http://szeliski.org/Book/. Suitable for an upper-level undergraduate or graduate-level course in computer science or engineering, this textbook focuses on basic techniques that work under real-world conditions and encourages students to push their creative boundaries. Its design and exposition also make it eminently suitable as a unique reference to the fundamental techniques and current research literature in computer vision.

/pdf/computer-vision-algorithms-and-applications-25dn6wu83j.pdf

Computer Vision: Algorithms and Applications

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Images containing faces are essential to intelligent vision-based human-computer interaction, and research efforts in face processing include face recognition, face tracking, pose estimation and expression recognition. However, many reported methods assume that the faces in an image or an image sequence have been identified and localized. To build fully automated systems that analyze the information contained in face images, robust and efficient face detection algorithms are required. Given a single image, the goal of face detection is to identify all image regions which contain a face, regardless of its 3D position, orientation and lighting conditions. Such a problem is challenging because faces are non-rigid and have a high degree of variability in size, shape, color and texture. Numerous techniques have been developed to detect faces in a single image, and the purpose of this paper is to categorize and evaluate these algorithms. We also discuss relevant issues such as data collection, evaluation metrics and benchmarking. After analyzing these algorithms and identifying their limitations, we conclude with several promising directions for future research.

/pdf/detecting-faces-in-images-a-survey-4yhbbns6mr.pdf

Detecting faces in images: a survey

In this work, we propose a computationally efficient compressive sensing based approach for very low bit rate lossy coding of hyperspectral (HS) image data by exploiting the redundancy inherent in this imaging modality. We divide the HS datacube into subsets of adjacent bands, each of which is encoded into a coded snapshot using a random code matrix. These coded snapshot images are encoded using the wavelet-based SPIHT compression technique. The decompression from the coded snapshots at the receiver is done using the orthogonal matching pursuit with the help of an overcomplete dictionary learned on a general purpose training dataset. We provide ample experimental results and performance comparisons to substantiate the usefulness of the proposed method. In the proposed technique the encoder is free from any decoder and it offers a significant saving in computation and yet yields a much higher compression quality.

A Fast Method for Very Low Bit-Rate Coding of Hyperspectral Images

Active metric learning is the problem of incrementally selecting high-utility batches of training data (typically, ordered triplets) to annotate, in order to progressively improve a learned model of a metric over some input domain as rapidly as possible Standard approaches, which independently assess the informativeness of each triplet in a batch, are susceptible to highly correlated batches with many redundant triplets and hence low overall utility While a recent work \cite{kumari2020batch} proposes batch-decorrelation strategies for metric learning, they rely on ad hoc heuristics to estimate the correlation between two triplets at a time We present a novel batch active metric learning method that leverages the Maximum Entropy Principle to learn the least biased estimate of triplet distribution for a given set of prior constraints To avoid redundancy between triplets, our method collectively selects batches with maximum joint entropy, which simultaneously captures both informativeness and diversity We take advantage of the submodularity of the joint entropy function to construct a tractable solution using an efficient greedy algorithm based on Gram-Schmidt orthogonalization that is provably $\left( 1 - \frac{1}{e} \right)$-optimal Our approach is the first batch active metric learning method to define a unified score that balances informativeness and diversity for an entire batch of triplets Experiments with several real-world datasets demonstrate that our algorithm is robust, generalizes well to different applications and input modalities, and consistently outperforms the state-of-the-art

A Unified Batch Selection Policy for Active Metric Learning

Having shown in previous chapters that joint MAP estimator does work for blind deconvolution, we proceed by selecting TV based priors for both the image as well as the PSF. Since both the image and the PSF are unknowns, alternate minimization (AM) is used to solve the blind deconvolution problem. In this chapter we provide a Fourier domain convergence analysis of the AM procedure. TV prior being non-linear, a non-quadratic cost function is obtained at each iteration of the AM. This makes the convergence analysis difficult. For making the analysis feasible, the TV prior is replaced by its quadratic upper bound. With this approximation, the cost function at each step of AM becomes quadratic in nature. Even with this approximation, the system obtained is not a linear shift invariant (LSI) system, which makes analysis in the spectral domain difficult. To overcome this, we use a further approximation, which makes the system LSI at each iteration – with the system changing with iteration. We note that the resulting system behaves like an adaptive Wiener filter. Once the fixed-point is reached the regularization factors remain constant, and a Fourier domain analysis shows that the fixed-points for image and PSF magnitudes are similar to those derived in an already existing analysis. Our analysis differs in that the image and the PSF regularization factors become signal (image/PSF) dependent. It is observed that the convergence points of the magnitudes of the image and the PSF are related to each other, which is due to the approximation to make the system LSI. We analyze the error term in the approximation, which provides an insight into the regularizing capability of the TV prior.

Convergence Analysis in Fourier Domain

This book presents and analyzes methods to perform image co-segmentation and includes several block diagrams and illustrative examples 

Image Co-segmentation

Recovery of rotational parameters is a very important problem in a navigation system. A constrained least squares (CLS) method where the estimated rotation matrix is constrained to be orthonormal or a quaternion based least squares method is used to estimate the parameters. However, the solution is restricted to analyzing the motion between two successive frames only. In this paper, we propose a computationally efficient recursive scheme to estimate the rotation parameters from a long sequence of data using the quaternion formulation. The proposed method is based on the rotation of the input data at different frames to approximately align them to a given subspace. The left (the right) singular subspace defined by the running sum of the outer products of the available data points is used as the alignment matrix. The exact alignment is then obtained using the singular value decomposition (SVD) technique. The recursive estimate of the quaternions representing the constant, fixed axis rotation is shown to be given by the vector corresponding to the smallest singular value. Thus, the proposed formulation is identical to the problem of recursive updating of symmetric matrices. A recursive CLS method is also developed using a similar technique, which would yield identical estimates as in the previous method, for a given data set. The details of these methods and simulation results are presented in this paper.

Subhasis Chaudhuri

Papers

A Fast Method for Very Low Bit-Rate Coding of Hyperspectral Images

A Unified Batch Selection Policy for Active Metric Learning

Convergence Analysis in Fourier Domain

Image Co-segmentation

Recursive solution to quaternion-based navigation