Background . Continuous wavelet transform (CWT) based scalogram can be used for photoplethysmography (PPG) signal transformation to classify blood pressure (BP) with deep learning. We aimed to investigate the determinants that can improve the accuracy of BP classification based on PPG and deep learning and establish a better algorithm for the prediction. Methods . The dataset from PhysioNet was accessed to extract raw PPG signals for testing and its corresponding BPs as category labels. The BP category of normal or abnormal followed the criteria of the 2017 American College of Cardiology/American Heart Association (ACC/AHA) Hypertension Guidelines. The PPG signals were transformed into 224 ∗ 224 ∗ 3‐pixel scalogram via different CWTs and segment units. All of them are fed into different convolutional neural networks (CNN) for training and validation. The receiver‐operating characteristic and loss and accuracy curves were used to evaluate and compare the performance of different methods. Results . Both wavelet type and segment length could affect the accuracy, and Cgau1 wavelet and segment‐300 revealed the best performance (accuracy 90%) without obvious overfitting. This method performed better than previously reported MATLAB Morse wavelet transformed scalogram on both of our proposed CNN and CNN‐GoogLeNet. Conclusions . We have established a new algorithm with high accuracy to predict BP classification from PPG via matching of CWT type and segment length, which is a promising solution for rapid prediction of BP classification from real‐time processing of PPG signal on a wearable device.

Background: As an important part of the tongue, the tongue coating is closely associated with different disorders and has major diagnostic benefits. This study aims to construct a neural network model that can perform complex tongue coating segmentation. This addresses the issue of tongue coating segmentation in intelligent tongue diagnosis automation. Method: This work proposes an improved TransUNet to segment the tongue coating. We introduced a transformer as a self-attention mechanism to capture the semantic information in the high-level features of the encoder. At the same time, the subtraction feature pyramid (SFP) and visual regional enhancer (VRE) were constructed to minimize the redundant information transmitted by skip connections and improve the spatial detail information in the low-level features of the encoder. Results: Comparative and ablation experimental findings indicate that our model has an accuracy of 96.36%, a precision of 96.26%, a dice of 96.76%, a recall of 97.43%, and an IoU of 93.81%. Unlike the reference model, our model achieves the best segmentation effect. Conclusion: The improved TransUNet proposed here can achieve precise segmentation of complex tongue images. This provides an effective technique for the automatic extraction in images of the tongue coating, contributing to the automation and accuracy of tongue diagnosis.

Convolutional Neural Networks (CNNs) dominate medical image classification, yet their “black box” nature limits understanding of their decision-making process. This study applies quantitative interpretability metrics to evaluate CNN performance in stained tongue coating recognition and compare with traditional metrics. We trained four classical CNN models (ResNet18, ResNet50, VGG19, and AlexNet) on a dataset of 2,008 tongue coating images, with external validation on 381 new images. Class Activation Mapping (CAM) algorithms generated heatmaps visualizing influential regions. The Heatmap Assisted Accuracy Score (HAAS) was utilized to assess feature attribution quality. All models achieved high classification performance on the test set (accuracy >0.92, precision >0.89, recall >0.91), but VGG19 and AlexNet performed poorly on external validation. Interpretability analysis revealed that VGG19 and AlexNet deviated from regions of interest, while ResNet models achieved significantly higher HAAS scores. ResNet50 emerged as the best model in external validation (accuracy =0.90, precision =0.863, recall = 0.911 ), consistent with its superior interpretability metrics (Eigen-CAM HAAS=1.551). Our findings demonstrate that interpretability metrics more accurately reflect CNN performance before external validation, offering valuable tools for understanding model behavior and enhancing transparency in medical image classification.