Machine intelligence and signal processing

Chapter 1. Advancing Cross-spectral Iris Recognition Research using Bi-spectral Imaging -- Chapter 2. Fast 3D Salient Region Detection in Medical Images using GPUs -- Chapter 3. Recovering Partially Sampled EEG Signals using Learned Dictionaries -- Chapter 4. Greedy Algorithms for Non-linear Sparse Recovery -- Chapter 5. Improving Rating Predictions by Baseline Estimation and Single Pass Low-rank Approximation -- Chapter 6. Reducing Inter-scanner Variability in Multi-site fMRI Activations using Correction Functions: A Preliminary Study -- Chapter 7. Genetically Modified Logistic Regression with Radial Basis Function for Robust Software Effort Prediction -- Chapter 8. Missing Data Interpolation using Compressive Sensing: An Application for Sales Data Gathering -- Chapter 9. Retinal Vessel Classification based on Maximization of Squared-loss Mutual Information -- Chapter 10. Retinal Blood Vessel Extraction and Optic Disc Removal using Curvelet Transform and Morphological Operation -- Chapter 11. Adaptive Skin Color Model to Improve Video Face Detection -- Chapter 12. Automated Spam Detection in Short Text Messages -- Chapter 13. Domain Adaptation for Face Detection -- Chapter 14. Comparative Study of Pre-processing and Classification Methods in Character Recognition of Natural Scene Images.

This book comprises chapters on key problems in machine learning and signal processing arenas. The contents of the book are a result of a 2014 Workshop on Machine Intelligence and Signal Processing held at the Indraprastha Institute of Information Technology. Traditionally, signal processing and machine learning were considered to be separate areas of research. However in recent times the two communities are getting closer. In a very abstract fashion, signal processing is the study of operator design. The contributions of signal processing had been to device operators for restoration, compression, et cetera Applied Mathematicians were more interested in operator analysis. Nowadays signal processing research is gravitating towards operator learning - instead of designing operators based on heuristics (for example wavelets), the trend is to learn these operators (for example dictionary learning). And thus, the gap between signal processing and machine learning is fast converging. The 2014 Workshop on Machine Intelligence and Signal Processing was one of the few unique events that are focused on the convergence of the two fields. The book is comprised of chapters based on the top presentations at the workshop. This book has three chapters on various topics of biometrics - two are on face detection and one on iris recognition; all from top researchers in their field. There are four chapters on different biomedical signal / image processing problems. Two of these are on retinal vessel classification and extraction; one on biomedical signal acquisition and the fourth one on region detection. There are three chapters on data analysis - a topic gaining immense popularity in industry and academia. One of these shows a novel use of compressed sensing in missing sales data interpolation. Another chapter is on spam detection and the third one is on simple one-shot movie rating prediction. Four other chapters cover various cutting edge miscellaneous topics on character recognition, software effort prediction, speech recognition and non-linear sparse recovery. The contents of this book will prove useful to researchers, professionals and students in the domains of machine learning and signal processing.