Ensemble of Neural Networks for Automated Cell Phenotype Image Classification

Subcellular location is related to the knowledge of the spatial distribution of a protein within the cell. The knowledge of the location of all proteins is crucial for several applications ranging from early diagnosis of a disease to monitoring of therapeutic effectiveness of drugs. This chapter focuses on the study of machine learning techniques for cell phenotype image classification and is aimed at pointing out some of the advantages of using a multi-classifier system instead of a stand-alone method to solve this difficult classification problem. The main problems and solutions proposed in this field are discussed and a new approach is proposed based on ensemble of neural networks trained by local and global features. Finally, the most used benchmarks for this problem are presented and an experimental comparison among several state-of-the-art approaches is reported which allows to quantify the performance improvement obtained by the approach proposed in this chapter.

Content-Based Access to Medical Image Collections

Medical images are a very important resource for the clinical practice and operation. Thousands of them are daily acquired in hospitals to diagnose the health state of patients. However, once they have been archived as part of a large image database, it is very difficult to retrieve the same image again since it requires remembering dates or names. Furthermore, in such cases in which the same image is not required, but a physician is looking for images with particular contents, current technologies are not able to offer such functionality. The ability to find the right visual information in the right place, at the right time, can have great impact in the medical decision making process. This chapter presents two computational strategies for accessing a large collection of medical images: retrieving relevant images given an explicit query and visualizing the structure of the whole collection. Both strategies take advantage of image contents, allowing users to find or identify images that are related by their visual composition. In addition, these strategies are based on machine learning methods to handle complex image patterns, semantic medical concepts, image collection visualizations and summarizations.

Left Ventricle Segmentation and Motion Analysis in MultiSlice Computerized Tomography

Cardiac motion analysis is an important tool for evaluating the cardiac function. Accurate motion estimation techniques are necessary for providing a set of parameters useful for diagnosis and guiding therapeutical actions. In this chapter, the problem of cardiac motion estimation is presented. A short overview of techniques based in several imaging modalities is given where the machine learning techniques have played an important role. A feasible solution for left ventricle segmentation in multislice computerized tomography (MSCT) and for estimating the left ventricle motion is presented. This method is based on the application of support vector machines (SVM), region growing and a nonrigid bidimensional correspondence algorithm used for tracking the anatomical landmarks extracted from the segmented left ventricle (LV). Some experimental results are presented and at the end of the chapter a short summary is presented.

Biomedical Microscopic Image Processing by Graphs

The authors present an overview of part of their work on graph-based regularization. Introduced first in order to smooth and filter images, the authors have extended these methods to address semi-supervised clustering and segmentation of any discrete domain that can be represented by a graph of arbitrary structure. This framework unifies, within a same formulation, methods from machine learning and image processing communities. In this chapter, the authors propose to show how these graph-based approaches can lead to a useful set of tools that can be combined altogether to address various image processing problems in pathology such as cytological and histological image filtering, segmentation and classification.

Predicting Complex Patterns of Human Movements Using Bayesian Online Learning in Medical Imaging Applications

Medical images, in its tough sense, are fundamental in most clinical procedures and have become part of the medical act. Different acquisition methodologies result in a large variety of challenges or diagnostic tasks. Overall, most applications are dedicated to imaging structures so that complex measurements may be achieved. However, function analysis necessitates imaging structures through the time, either at the level of the image itself or at the interaction strategy between the user and the image. This chapter presents a Bayesian Framework which allows an adequate temporal follow up of very complex human movements, which somehow have been imaged. The Bayesian strategy is implemented through a particle filter, resulting in real time tracking of these complex patterns. Two different imaged patterns illustrate the potential of the procedure: a precise tracking a pathologist in a Virtual Microscopy context and a temporal follow up of gait patterns.

Machine Learning for Brain Image Segmentation

In this chapter, the authors review a variety of algorithms developed by different groups for automatically segmenting structures in medical images, such as brain MRI scans. Some of the simpler methods, based on active contours, deformable image registration, and anisotropic Markov random fields, have known weaknesses, which can be largely overcome by learning methods that better encode knowledge on anatomical variability. The authors show how the anatomical segmentation problem may be re-cast in a Bayesian framework. They then present several different learning techniques increasing in complexity until they derive two algorithms recently proposed by the authors. The authors show how these automated algorithms are validated empirically, by comparison with segmentations by experts, which serve as independent ground truth, and in terms of their power to detect disease effects in Alzheimer’s disease. They show how these methods can be used to investigate factors that influence disease progression in databases of thousands of images. Finally the authors indicate some promising directions for future work.

A Genetic Algorithm-Based Level Set Curve Evolution for Prostate Segmentation on Pelvic CT and MRI Images

A novel genetic algorithm (GA) is presented here that performs level set curve evolution using texture and shape information to automatically segment the prostate on pelvic images in computed tomography and magnetic resonance imaging modalities. Here, the segmenting contour is represented as a level set function. The contours in a typical level set evolution are deformed by minimizing an energy function using the gradient descent method. In these methods, the computational complexity of computing derivatives increases as the number of terms (needed for curve evolution) in the energy function increase. In contrast, a genetic algorithm optimizes the level-set function without the need to compute derivatives, thereby making the introduction of new curve evolution terms straightforward. The GA developed here uses the texture of the prostate gland and its shape derived from manual segmentations to perform curve evolution. Using these high-level features makes automatic segmentation possible.

Computer-Aided Detection and Diagnosis of Breast Cancer Using Machine Learning, Texture and Shape Features

Breast cancer is a malignant (cancer) tumor that starts from cells of the breast, being the major cause of deaths by cancer in the female population. There has been tremendous interest in the use of image processing and analysis techniques for computer aided detection (CAD)/ diagnostics (CADx) in digital mammograms. The goal has been to increase diagnostic accuracy as well as the reproducibility of mammographic interpretation. CAD/CADx systems can aid radiologists by providing a second opinion and may be used in the first stage of examination in the near future, providing the reduction of the variability among radiologists in the interpretation of mammograms. This chapter provides an overview of techniques used in computer-aided detection and diagnosis of breast cancer. The authors focus on the application of texture and shape tissues signature used with machine learning techniques, like support vector machines (SVM) and growing neural gas (GNG).

Variational Approach Based Image Pre- Processing Techniques for Virtual Colonoscopy

Colorectal cancer includes cancer of the colon, rectum, anus and appendix. Since it is largely preventable, it is extremely important to detect and treat the colorectal cancer in the earliest stage. Virtual colonoscopy is an emerging screening technique for colon cancer. One component of virtual colonoscopy, image preprocessing, is important for colonic polyp detection/diagnosis, feature extraction and classification. This chapter aims at an accurate and fast colon segmentation algorithm and a general variational-approach based framework for image pre-processing techniques, which include 3D colon isosurface generation and 3D centerline extraction for navigation. The proposed framework has been validated on 20 real CT Colonography (CTC) datasets. The average segmentation accuracy has achieved 96.06%, and it just takes about 5 minutes for a single CT scan of 512*512*440. All the 12 colonic polyps with sizes of 6 mm and above in the 20 clinical CTC datasets are found by this work.

Genetic Adaptation of Level Sets Parameters for Medical Imaging Segmentation

This chapter presents a method based on level sets to segment organs using computer tomography (CT) medical images. Initially, the organ boundary is manually set in one slice as an initial solution, and then the method automatically segments the organ in all other slices, sequentially. In each step of iteration it fits a Gaussian curve to the organ’s slice histogram to model the speed image in which the level sets propagate. The parameters of our method are estimated using genetic algorithms (GA) and a database of reference segmentations. The method was tested to segment the liver using 20 different exams and five different measures of performance, and the results obtained confirm the potential of the method. The cases in which the method presented a poor performance are also discussed in order to instigate further research.