Share this article:
Biomedical Texture Analysis: Fundamentals, Tools and Challenges
Computerized recognition and quantification of texture information has been an active research domain for the past 50 years, with some of the pioneering work still widely used today. Recently, the increasing ubiquity of imaging data has driven the need for powerful image analysis approaches to convert this data into knowledge. One of the most promising application domains is biomedical imaging, which is a key enabling technology for precision medicine (e.g., radiomics and digital histopathology) and biomedical discovery (e.g., microscopy). The colossal research efforts and progress made in the general domain of computer vision have led to extremely powerful data analysis systems. Biomedical imaging relies upon well-defined acquisition protocols to produce images. This is quite different from general photography. Consequently, the analysis of biomedical images requires a paradigm change to account for the quantitative nature of the imaging process. Texture analysis is a broadly applicable, powerful technology for quantitative analysis of biomedical images.
The aims of the book are:
– Define biomedical texture precisely and describe how it is different from general texture information considered in computer vision;
– Define the general problem to translate 2D and 3D texture patterns from biomedical images to visually and biologically relevant measurements;
– Describe with intuitive concepts how the most popular biomedical texture analysis approaches (e.g., gray-level matrices, fractals, wavelets, deep convolutional neural networks) work, what they have in common, and how they are different;
– Identify the strengths, weaknesses, and current challenges of existing methods including both handcrafted and learned representations, as well as deep learning. The goal is to establish foundations for building the next generation of biomedical texture operators;
– Showcase applications where biomedical texture analysis has succeeded and failed;
– Provide details on existing, freely available texture analysis software. This will help experts in medicine or biology develop and test precise research hypothesis.
The book provides a thorough background on texture analysis for graduate students, and biomedical engineers from both industry and academia who have basic image processing knowledge. Medical doctors and biologists with no background in image processing will also find available methods and software tools for analyzing textures in medical images.
By bringing together experts in data science, medicine, and biology, we hope that this book will actively promote the translation of incredibly powerful data analysis methods into several breakthroughs in biomedical discovery and noninvasive precision medicine.
For a limited time you can read Chapter 1: Fundamentals of Texture Processing for Biomedical Image Analysis: A General Definition and Problem Formulation on ScienceDirect.
This chapter aims to provide an overview of the foundations of texture processing for biomedical image analysis. Its purpose is to define precisely what biomedical texture is, how is it different from general texture information considered in computer vision, and what is the general problem formulation to translate 2D and 3D textured patterns from biomedical images to visually and biologically relevant measurements. First, a formal definition of biomedical texture information is proposed from both perceptual and mathematical point of views. Second, a general problem formulation for biomedical texture analysis is introduced, considering that any approach can be characterized as a set of local texture operators and regional aggregation functions. The operators allow locally isolating desired texture information in terms of spatial scales and directions of a texture image. The type of desirable operator invariances are discussed, and are found to be different from photographic image analysis. Scalar-valued texture measurements are obtained by aggregating operator’s response maps over regions of interest.
The highly interdisciplinary field of materials science examines elements of applied physics and chemistry, as well as chemical, mechanical, civil, and electrical engineering. Nanoscience and nanotechnology in particular have yielded major innovations in this area, such as graphene and carbon nanotubes. Elsevier’s authoritative content in this area ranges from undergraduate textbooks to multi-volume reference works investigating the relationships between the structure of materials and their properties. Our journals (including Materials Today), books, and eBooks help researchers stay abreast of developments in this swiftly advancing field, coving major sub-disciplines like energy and power; metals and alloys; ceramics; composite materials; polymer science and biomaterials; interdisciplinary materials science; and structural materials.