RESUMO
The sustainable use of water resources is inseparable from water pollution detection. The sensing of toxic ammonia nitrogen in water currently requires auxiliary reagents, which may cause secondary pollution. Benefiting from the ability of substances to change light characteristics, this work proposes polarimetry-inspired feature fusion spectroscopy (PIFFS) to detect ammonia. The PIFFS system mainly includes a light source, a quarter-wave plate (QWP), a linear polarizer (LP) and a fiber spectrometer. The target light containing substance information is polarization modulated by adjusting the QWP and LP angles. Then, the Stokes parameters of target light can be calculated by appropriate modulations. The feasibility of PIFFS method to detect ammonia nitrogen is verified by experiments on both standard water samples and environmental water samples. Experimental results show that inspired by the first Stokes parameter, the fused features provide superiority in classifying ammonia concentration. The results also demonstrate the effectiveness of support vector machine-based concentration classification and random forests-based spectral selection. The interaction between light and substances ensures that the proposed PIFFS method has the potential to detect other pollutants.
RESUMO
BACKGROUND: Although many machine learning algorithms have been developed to detect anterior cruciate ligament (ACL) injury based on magnetic resonance imaging (MRI), the performance of different algorithms required further investigation. The objectives of this current systematic review are to evaluate the diagnostic accuracy of machine-learning-assisted detection for ACL injury based on MRI and find the current best algorithm. METHOD: We will conduct a comprehensive database search for clinical diagnostic tests in PubMed, EMBASE, Cochrane Library, and Web of science without restrictions on publication status and language. The reference lists of the included articles will also be checked to identify additional studies for potential inclusion. Two reviewers will independently review all literature for inclusion and assess their methodological quality using Quality Assessment of Diagnostic Accuracy Studies version 2. Clinical diagnostic tests exploring the efficacy of machine-learning-assisted system for detecting ACL injury based on MRI will be considered for inclusion. Another 2 reviewers will independently extract data from eligible studies based on a pre-designed standardized form. Any disagreements will be resolved by consensus. RevMan 5.3 and Stata SE 12.0 software will be used for data synthesis. If appropriate, we will calculate the summary sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio of machine-learning-assisted diagnosis system for ACL injury detection. A hierarchical summary receiver operating characteristic (HSROC) curve will also be plotted, and the area under the ROC curve (AUC) is going to calculated using the bivariate model. If the pooling of results is considered inappropriate, we will present and describe our findings in diagrams and tables and describe them narratively. RESULT: This is the first systematic assessment of machine learning system for the detection of ACL injury based on MRI. We predict it will provide highquality synthesis of existing evidence for the diagnostic accuracy of machine-learning-assisted detection for ACL injury and a relatively comprehensive reference for clinical practice and development of interdisciplinary field of artificial intelligence and medicine. CONCLUSION: This protocol outlined the significance and methodologically details of a systematic review of machine-learning-assisted detection for ACL injury based on MRI. The ongoing systematic review will provide high-quality synthesis of current evidence of machine learning system for detecting ACL injury. REGISTRATION: The meta-analysis has been prospectively registered in PROSPERO (CRD42019136581).
