Deep Learning in Healthcare

By: D. Peraković , K. Yadav, C. Hsu

In recent times, Deep learning has brought a revolutionary development in the field of healthcare and medicine. The ability of deep learning to abstract high-level knowledge from complex data and the generation of sophisticated features sets from data without any human intervention have highly contributed to the field of health science [1-4]. Deep learning is inspired by biological neurons like perceptron, which is activated when the information from other perceptrons injects it. A large number of the perceptron, when stacked together, forms a hidden layer [5, 6]. Numerous amounts of the hidden layer are used to process complex problems and obtain a sophisticated decision from a large number of datasets (Figure 1).

Application of Deep Learning in Healthcare

Deep learning methods are being used to explore the structure and information present in the DNA sequence of living cells. This information, when processed, identifies gene alleles that are contributing to diseases in a human being. Identification and study of these alleles then help to develop medical therapies to fight against these diseases [7, 8]. The Convolution neural network (CNN) has highly contributed to the medical Imaging of neurological diseases in human beings such as Alzheimer’s and stroke, which require the image scanning of different parts of the brain. To do medical Imaging, a large number of medical images are required, which is a barrier; however, manual annotations are being done by several experts to develop a large number of datasets. CNN has been beneficial in analyzing 2-D images; however, some medical process such as MRI produces 3-D images which are challenging to process.

Data Pre-processing for Deep Learning in Healthcare

Several isotropic images are being developed from 3-D images so that CNN can process them; however, it highly suffers from blurriness. Research is being carried out for the development of high-quality isotropic images by reducing the dimension of high dimensional images. Several wearable devices with sensors are being developed to collect information regarding person movements, calorie intake, and blood pressure, which are then processed by several deep learning algorithms. The result obtained from these algorithms then notifies people about their present health conditions. To provide telepathy medical services, it is necessary to develop an affordable wearable sensor device powered up by highly optimized deep learning algorithms.

Deep learning for COVID-19 Detection

In 2019 a disease that causes lung illness started in China, and in no time become a worldwide epidemic. As there is no vaccine available for this disease it affects a large number of populations of the world. However, COVID-19 has some symptoms that help the research for quick identification of COVID-19 patients. In this context, Deep Learning algorithms are showing good results in the detection of COVID-19 cases at the early stage [1].  Recently, Sedik et al. [8] proposed a novel COVID-19 detection technique based on a convolutional neural network and ConvLSTM. In the proposed approach, the Deep Learning algorithms compares the chest X-Ray of normal patients with the COVID-19 patients, as represented in Figure 2. The proposed gives the accuracy of 99%

Research Arears for Deep Learning in Healthcare

Some of the eminent research areas on healthcare based on Deep Learning are following [9, 10]:

• Data mining techniques in Bioinformatics
• Medical Imaging
• High dimensional Image processing
• Development of artificial medical datasets
• Convolution neural network
• Gene’s classifications
• Low-cost medical sensor development
• Development of medical drugs
• Machine learning-based cancer classification
• Anomaly detection in medical Images
• Human activity tracker
• Environmental pollutions prediction
• Gesture recognition
• Disease’s visualization
• Medical Image segmentation
• Automation in diseases classification
• Brain tumour detection
• Natural language processing in medical query management
• Physical abnormalities analysis with CNN
• Deep learning for drug analysis

Keywords: Convolution neural network, Health, Images, Datasets, Sensors, Diseases, Medical Imaging, Perceptron, Medicine

References

1. M. S. Hossain and G. Muhammad. 2020. Deep learning-based pathology detection for smart connected healthcares. IEEE Network. 34, 6 (2020), 120–125.
2. Ravì, D., Wong, C., Deligianni, F., Berthelot, M., Andreu-Perez, J., Lo, B., & Yang, G. Z. (2016). Deep learning for health informatics. IEEE journal of biomedical and health informatics, 21(1), 4-21.
3. Havaei, M., Guizard, N., Larochelle, H., & Jodoin, P. M. (2016). Deep learning trends for focal brain pathology segmentation in MRI. In Machine learning for health informatics (pp. 125-148). Springer, Cham.
4. Zhao, R., Yan, R., Chen, Z., Mao, K., Wang, P., & Gao, R. X. (2019). Deep learning and its applications to machine health monitoring. Mechanical Systems and Signal Processing, 115, 213-237.
5. R. Mehla (2021). Application of Deep Learning in Big Data Analytics for Healthcare Systems, Insights2Techinfo, pp. 1
6. Srivastava, S., Soman, S., Rai, A., & Srivastava, P. K. (2017, September). Deep learning for health informatics: Recent trends and future directions. In 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI) (pp. 1665-1670). IEEE.
7. Jyotiyana, M., & Kesswani, N. (2021). Introduction to Deep Learning in Health Informatics. Biomedical Data Mining for Information Retrieval: Methodologies, Techniques and Applications, 237-261.
8. Sedik, A., Hammad, M., Abd El-Samie, F. E., et al. (2021). Efficient deep learning approach for augmented detection of Coronavirus disease. Neural Computing and Applications, 1-18.
9. Yadav, K., Quamara, M., & Gupta, B. 2021 Hot Topics in Machine Learning Research. Insights2Techinfo, pp.1
10. Chai, Y., Bian, Y., Liu, H., Li, J., & Xu, J. (2021). Glaucoma diagnosis in the Chinese context: An uncertainty information-centric Bayesian deep learning model. Information Processing & Management, 58(2), 102454.

Cite this article

D. Peraković, K. Yadav, C. Hsu (2021) Deep Learning in Healthcare, Insights2Techinfo, pp.1

FAQ on this topic

118990cookie-checkDeep Learning in Healthcareyes