Abstract
Tongue color is an important part of tongue diagnosis. The change of tongue color is affected by pathological state of body, blood rheology, and other factors. Therefore, physicians can understand a patient’s condition by observing tongue color. Currently, most studies use machine learning, which is time consuming and labor intensive. Other studies use deep learning based on convolutional neural network (CNN), but the affine transformation of CNN is less robust and easily loses the spatial relationship between features. Recently, Capsule Networks (CapsNet) have been proposed to overcome these problems. In our work, CapsNet is used for tongue color research for the first time, and improved model TongueCaps is proposed, which combines the advantage of CapsNet and residual block structure to achieve end to end tongue color classification. We conduct experiments on 1371 tongue images; TongueCaps achieve accuracy is 0.8456, sensitivity is 0.8474, and specificity is 0.9586. In addition, the size of TongueCaps is 8.11 M, and FLOPs is 1,335,342, which are smaller than CNN in comparison models. Experiments have confirmed that the CapsNet can be used for tongue color research, and improved model TongueCaps, in this paper, is superior to other comparison models in terms of accuracy, specificity and sensitivity, computational complexity, and size of model.
Highlights
Tongue diagnosis has been recorded as early as in the classic Chinese medicine book ‘Huangdi Neijing’, which was to diagnose disease by observing the characteristics of tongue, and it made rapid progress in the diagnosis of exogenous fever
We proposed TongueCaps based on Capsule Networks (CapsNet) and compared the performance of this model with other common convolutional neural network (CNN) models on the tongue color classification task through experiments
Our model has been improved in the feature extraction stage, increasing the number of convolutional layers and introducing a residual block structure to alleviate the over fitting phenomenon and enhance the model’s ability to extract features
Summary
Tongue diagnosis has been recorded as early as in the classic Chinese medicine book ‘Huangdi Neijing’, which was to diagnose disease by observing the characteristics of tongue, and it made rapid progress in the diagnosis of exogenous fever. Tongue diagnosis has become a unique diagnostic method under the guidance of Traditional Chinese Medicine (TCM) theory [1]. The content of tongue examination is divided into two parts: observation of tongue texture and tongue coating. Tongue color is an important content of tongue texture, and is generally divided into five categories: light red, red, deep red, light white, cyan [2]. Different colors can reflect different physiological and pathological states, blood rheology [2], and the attributes of pathogens [3], which is an important basis for effective clinical diagnosis, guidance of medication, efficacy judgment, and prognosis [2,4]. Correct identification of tongue color is of great clinical significance
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have