Deep-DSP: deep convolutional network with double spatial pyramid for tongue image segmentation
by Yuhan Chen; Qingfeng Li; Dongxin Lu; Wende Ke; Wenming Cui; Lei Kou
International Journal of Bio-Inspired Computation (IJBIC), Vol. 23, No. 3, 2024

Abstract: In traditional Chinese medicine diagnosis, experienced Chinese medicine doctors can understand the potential symptoms of patients by diagnosing their tongue images. With the continuous development of computer science and artificial intelligence, neural networks are gradually being used in the image processing and feature extraction. The intelligent tongue diagnosis combining traditional Chinese medicine with neural networks was constructed to automatically improve the diagnose accuracy. The dataset contains 556 no-background tongue images. The deep convolutional network with double spatial pyramid (Deep-DSP) for tongue segmentation was constructed, in which the spatial pyramid block was embedded into the encoder-decoder architecture. Noteworthy, a novel double spatial pyramid architecture was applied for multiple encoding states feature learning. Experiments showed that the deep-DSP performed more excellently in tongue segmentation and than that of state-of-the-art methods according to the accuracy, precision, recall, harmonic measure and specificity.

Online publication date: Mon, 08-Apr-2024

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Bio-Inspired Computation (IJBIC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com