Survey of Transfer Learning Approaches in the Machine Learning of Digital Health Sensing Data.

Machine learning and digital health sensing data have led to numerous research achievements aimed at improving digital health technology. However, using machine learning in digital health poses challenges related to data availability, such as incomplete, unstructured, and fragmented data, as well as issues related to data privacy, security, and data format standardization. Furthermore, there is a risk of bias and discrimination in machine learning models. Thus, developing an accurate prediction model from scratch can be an expensive and complicated task that often requires extensive experiments and complex computations. Transfer learning methods have emerged as a feasible solution to address these issues by transferring knowledge from a previously trained task to develop high-performance prediction models for a new task. This survey paper provides a comprehensive study of the effectiveness of transfer learning for digital health applications to enhance the accuracy and efficiency of diagnoses and prognoses, as well as to improve healthcare services. The first part of this survey paper presents and discusses the most common digital health sensing technologies as valuable data resources for machine learning applications, including transfer learning. The second part discusses the meaning of transfer learning, clarifying the categories and types of knowledge transfer. It also explains transfer learning methods and strategies, and their role in addressing the challenges in developing accurate machine learning models, specifically on digital health sensing data. These methods include feature extraction, fine-tuning, domain adaptation, multitask learning, federated learning, and few-/single-/zero-shot learning. This survey paper highlights the key features of each transfer learning method and strategy, and discusses the limitations and challenges of using transfer learning for digital health applications. Overall, this paper is a comprehensive survey of transfer learning methods on digital health sensing data which aims to inspire researchers to gain knowledge of transfer learning approaches and their applications in digital health, enhance the current transfer learning approaches in digital health, develop new transfer learning strategies to overcome the current limitations, and apply them to a variety of digital health technologies.

DeepQuantify: deep learning and quantification system of white blood cells in light microscopy images of injured skeletal muscles.

White blood cells (WBCs) are the most diverse types of cells observed in the healing process of injured skeletal muscles. In the recovery process, WBCs exhibit a dynamic cellular response and undergo multiple changes of the protein expression. The progress of healing can be analyzed by the number of WBCs or by the number of specific proteins observed in light microscopy images obtained at different time points after injury. We propose a deep learning quantification and analysis system called DeepQuantify to analyze WBCs in light microscopy images of uninjured and injured muscles of female mice. The DeepQuantify system features in segmentation using the localized iterative Otsu's thresholding method, masking postprocessing, and classification of WBCs with a convolutional neural network (CNN) classifier to achieve a high accuracy and a low manpower cost. The proposed two-layer CNN classifier designed based on the optimization hypothesis is evaluated and compared with other CNN classifiers. The DeepQuantify system adopting these CNN classifiers is evaluated for quantifying CD68-positive macrophages and 7/4-positive neutrophils and compared with the state-of-the-art deep learning segmentation architectures. DeepQuantify achieves an accuracy of 90.64% and 89.31% for CD68-positive macrophages and 7/4-positive neutrophils, respectively. The DeepQuantify system employing the proposed two-layer CNN architecture achieves better performance than those deep segmentation architectures. The quantitative analysis of two protein dynamics during muscle recovery is also presented.

Sparse-view neutron-photon computed tomography: Object reconstruction and material discrimination.

Taking into account the advantages of both neutron- and photon-based systems, we propose combined neutron-photon computed tomography (CT) under a sparse-view setting and demonstrate its performance for 3D object visualization and material discrimination. We use a high-performance regularization method for CT reconstruction by combining regularization based on total variation (TV) and curvelet transform in cone beam geometry. It is coupled with proposed 2D material signatures which is pairs of photon to neutron transmission ratios and neutron transmission values per object space voxels. Classification of materials is performed by association of a voxel signature with library signatures; and per object - by majority of voxels in the object. Representation of object-material pairs, for the model in our experiment, a complex scene with group of high-Z and low-Z materials, attains the reconstruction accuracy of 92.1% and the overall high-Z discrimination accuracy of object representation is 85%, and by about 7.5% higher discrimination accuracy than that with 1D signatures which are ratios of photon to neutron transmissions. With a relative noise level of 10%, the method yields the reconstruction accuracies of 87.2%. The analyses are performed in cone beam configuration, with Monte Carlo modeling of neutron-photon transport for the model of object geometry and material contents.

Compressed sensing magnetic resonance imaging based on shearlet sparsity and nonlocal total variation.

Compressed sensing (CS) has been utilized for acceleration of data acquisition in magnetic resonance imaging (MRI). MR images can then be reconstructed with an undersampling rate significantly lower than that required by the Nyquist sampling criterion. However, the CS usually produces images with artifacts, especially at high reduction rates. We propose a CS MRI method called shearlet sparsity and nonlocal total variation (SS-NLTV) that exploits SS-NLTV regularization. The shearlet transform is an optimal sparsifying transform with excellent directional sensitivity compared with that by wavelet transform. The NLTV, on the other hand, extends the TV regularizer to a nonlocal variant that can preserve both textures and structures and produce sharper images. We have explored an approach of combining alternating direction method of multipliers (ADMM), splitting variables technique, and adaptive weighting to solve the formulated optimization problem. The proposed SS-NLTV method is evaluated experimentally and compared with the previously reported high-performance methods. Results demonstrate a significant improvement of compressed MR image reconstruction on four medical MRI datasets.

Microcalcification detection based on wavelet domain hidden markov tree model: study for inclusion to computer aided diagnostic prompting system.

In this paper we investigate the performance of statistical modeling of digital mammograms by means of wavelet domain hidden Markov trees for its inclusion to a computer-aided diagnostic prompting system. The system is designed for detecting clusters of microcalcifications. Their further discrimination as benign or malignant is to be done by radiologists. The model is used for segmenting images based on the maximum likelihood classifier enhanced by the weighting technique. Further classification incorporates spatial filtering for a single microcalcification (MC) and microcalcification cluster (MCC) detection. Contrast filtering applied for the digital database for screening mammography (DDSM) dataset prior to spatial filtering greatly improves the classification accuracy. For all MC clusters of 40 mammograms from the mini-MIAS database of Mammographic Image Analysis Society, 92.5%-100% of true positive cases can be detected under 2-3 false positives per image. For 150 cases of DDSM cases, the designed system is capable to detect up to 98% of true positives under 3.3% of false positive cases.

Detecting microcalcifications in digital mammograms using wavelet domain hidden Markov tree model.

In this paper we investigate the performance of statistical modeling of digital mammograms by means of wavelet domain hidden Markov tree model (WHMT) for its inclusion to a computer-aided diagnostic prompting system for detecting microcalcification (MC) clusters. The system incorporates: (1) gross-segmentation of mammograms for obtaining the breast region; (2) eliminating the pepper-type noise, (3) block-wise wavelet transform of the breast signal and likelihood calculation; (4) image segmentation; (5) postprocessing for retaining MC clusters. FROC curves are obtained for all MC clusters containing mammograms of mini-MIAS database. 100% of true positive cases are detected by the system at 2.9 false positives per case.