Machine learning techniques for biomedical image segmentation: An overview of technical aspects and introduction to state‐of‐art applications

Abstract

In recent years, significant progress has been made in developing more accurate and efficient machine learning algorithms for segmentation of medical and natural images. In this review article, we highlight the imperative role of machine learning algorithms in enabling efficient and accurate segmentation in the field of medical imaging. We specifically focus on several key studies pertaining to the application of machine learning methods to biomedical image segmentation. We review classical machine learning algorithms such as Markov random fields, k‐means clustering, random forest, etc. Although such classical learning models are often less accurate compared to the deep‐learning techniques, they are often more sample efficient and have a less complex structure. We also review different deep‐learning architectures, such as the artificial neural networks (ANNs), the convolutional neural networks (CNNs), and the recurrent neural networks (RNNs), and present the segmentation results attained by those learning models that were published in the past 3 yr. We highlight the successes and limitations of each machine learning paradigm. In addition, we discuss several challenges related to the training of different machine learning models, and we present some heuristics to address those challenges.

Auto‐segmentation of organs at risk for head and neck radiotherapy planning: From atlas‐based to deep learning methods

“…The underlying DL model is, in comparison to ABAS, also more robust because it can be trained with all available data, including patients with metal artifacts and diverse anatomy. 7 The main advantage of DL-based autosegmentation is in its ability to systematically learn the most adequate features for segmentation from a set of annotated training images, and then automatically search for the same features in a previously unseen image. Although this proved to result in the best overall segmentation performance, 49 it is not without drawbacks.…”

“…In the past decade, the field of computerized medical imaging has experienced an increased interest, with new emerging trends that are largely focused on deep learning (DL) 6 as a subset of machine learning that mimics the data processing of the human brain for the purpose of decision‐making. In comparison to traditional approaches based on conventional atlases, shape models and feature classification, DL has shown superior image segmentation performance that was conveyed by several milestone auto‐segmentation frameworks, 7 for example, the U‐Net, 8 3D U‐Net, 9 V‐Net, 10 SegNet, 11 DeepMedic, 12 DeepLab, 13 VoxResNet 14 and Mask R‐CNN 15 . Several ideas have been adopted for RT, 16,17 including for image segmentation and detection, image phenotyping, radiomic signature discovery, clinical outcome prediction, image dose quantification, dose‐response modeling, radiation adaptation, and image generation, 18 and therefore also impacted the area of auto‐segmentation of OARs in the H&N region 19–21 so as to provide a qualitative support for guiding critical treatment planning and delivery decisions.…”

Glass-cutting medical images via a mechanical image segmentation method based on crack propagation

“…Within systems lacking transparency, deep-learning networks require a great amount of hyperparameter tuning. Small changes in the hyperparameters can result in disproportionately large changes in the network output 16 . Moreover, there is a known problem with deep neural networks, where visually indistinguishable images can return significantly different results.…”

“…When computer models are used to simulate physiological phenomena, explore pathogenesis, and design personalized surgery, image segmentation is an essential step for reconstructing the anatomical structure of relevant tissues and organs [2][3][4] . Some typical segmentation technologies, such as the active contour model [5][6][7][8][9][10] , atlas-based registration [11][12][13][14] , and neural network-based segmentation [15][16][17][18] , have become more mature over the past several decades. Other strategies, such as fuzzy clustering 19 , the superpixel method 20,21 , and graph-cut method 22,23 , are also well applied to medical image segmentation.…”

Whole-body voxel-based internal dosimetry using deep learning

“…Yet, this approach is not employed in routine clinical procedures owing to the heavy computational burden. Deep learning emerged as a promising technique in the area of computer vision and image processing, exhibiting superior performance over conventional state-of-the-art methods in medical image analysis in PET and SPECT imaging, including attenuation and scatter correction [22][23][24], low-count image reconstruction [25][26][27], and automated image segmentation [9,28]. More recently, deep learning approaches were employed for radiation dose estimation.…”