Non-negative matrix factorization (NMF) controls negative bases in the principal component analysis (PCA) with non-negative constraints for basis and coefficient matrices. Semi-, convex-, and sparse-NMF modify these constraints to establish distinct properties for various applications in different fields, particularly in infrared thermography. In this study, we delve into the applications of semi-, convex-, and sparse-NMF in infrared diagnostic imaging systems. We applied these approaches to active and passive thermographic imaging systems to determine the heterogeneous thermal patterns in these sets. In active thermography, three diverse specimens, carbon fiber-reinforced polymer composites (CFRP), poly(methyl methacrylate) (PMMA, also known as Plexiglas), and aluminum plate, were used. Quantitative analyses were performed using the Jaccard index. In passive thermography, 55 participants for infrared breast screening selected from the Database for Mastology Research (DMR) dataset with symptomatic and healthy participants. We calculated five derived properties of the breast area (contrast, correlation, dissimilarity, homogeneous, and energy) by using thermal level co-occurrence matrices (TLCMs) and trained a logistic regression method to stratify between healthy and symptomatic patients. For both scenarios, we compared the ability of semi-, convex-, and sparse-NMF to state-of-the-art thermographic approaches, such as principal component analysis/thermography (PCT), candid covariance-free incremental principal component thermography (CCIPCT), sparse-PCT, and NMF. Measurement of different defect depths and sizes indicated significant performance for sparse-NMF (AL(d < 1mm):

Thermography has been used extensively as a complementary diagnostic tool in breast cancer detection. Among thermographic methods matrix factorization (MF) techniques show an unequivocal capability to detect thermal patterns corresponding to vasodilation in cancer cases. One of the biggest challenges in such techniques is selecting the best representation of the thermal basis. In this study, an embedding method is proposed to address this problem and Deep-semi-nonnegative matrix factorization (Deep-SemiNMF) for thermography is introduced, then tested for 208 breast cancer screening cases. First, we apply Deep-SemiNMF to infrared images to extract lowrank thermal representations for each case. Then, we embed lowrank bases to obtain one basis for each patient. After that, we extract 300 thermal imaging features, called thermomics, to decode imaging information for the automatic diagnostic model. We reduced the dimensionality of thermomics by spanning them onto Hilbert space using RBF kernel and select the three most efficient features using the block Hilbert Schmidt Independence Criterion Lasso (block HSIC Lasso). The preserved thermal heterogeneity successfully classified asymptomatic versus symptomatic patients applying a random forest model (crossvalidated accuracy of 71.36% (69.42%-73.3%)).

Detection of subsurface defects is undeniably a growing subfield of infrared non-destructive testing (IR-NDT). There are many algorithms used for this purpose, where non-negative matrix factorization (NMF) is considered to be an interesting alternative to principal component analysis (PCA) by having no negative basis in matrix decomposition. Here, an application of Semi non-negative matrix factorization (Semi-NMF) in IR-NDT is presented to determine the subsurface defects of an Aluminum plate specimen through active thermographic method. To benchmark, the defect detection accuracy and computational load of the Semi-NMF approach is compared to state-of-the-art thermography processing approaches such as: principal component thermography (PCT), Candid Covariance-Free Incremental Principal Component Thermography (CCIPCT), Sparse PCT, Sparse NMF and standard NMF with gradient descend (GD) and non-negative least square (NNLS). The results show 86% accuracy for 27.5s computational time for SemiNMF, which conclusively indicate the promising performance of the approach in the field of IR-NDT.

The presented approach addresses a review on the overheating which occurs during radiological examinations such as MRI and a series of thermal experiments to determine the thermal suitable fabric material which should be used for radiological gowns. Moreover, an automatic system for detecting and tracking of the thermal fluctuation is presented. It applies HSV based kernelled k-means clustering which initializes and controls the points which lie on the Region of Interest (ROI) boundary. Afterwards a particle filter tracks the targeted ROI during the video sequence independent to previous locations of the overheating spots. The proposed approach was tested during some experiments and under conditions very similar to those used during real radiology exams. Six subjects have voluntarily participated in these experiments. To simulate the hot spots occurring during the radiology, a controllable heat source was utilized near the subjects body. The results indicate promising accuracy for the proposed approach to track the hot spots. Some approximations were used regarding the * Bardia Yousefi, Xavier P.V. MaldagueEmail address: bardia.yousefi@ieee.org and Xavier.Maldague@gel.ulaval.ca. Tel: (+1)418-656-2962 (Bardia Yousefi, Julien Fleuret, Hai Zhang., Xavier P.V. Maldague ) Preprint submitted to Draft version of Applied OpticNovember 21, 2016 transmittance of the atmosphere and emissivity of the fabric could be neglected because of the independency of the proposed approach for these parameters.The approach can track the heating spots continuously and correctly, even for moving subjects, and provides considerable robustness against motion artifact, which usually occurs during most medical radiology procedures.