Abstract

Dermatology is the branch of bioscience that's involved with diagnosing and treatment of skin based mostly
disorders. The immense spectrum of dermatologic disorders varies geographically and additionally seasonally
because of temperature, humidness and alternative environmental factors. Human skin is one amongst the
foremost unpredictable and tough terrains to mechanically synthesize and analyze because of its quality of
unevenness, tone, presence of hair and alternative mitigating options. Though, many research are conducted
to find and model human skin victimisation (PC Vision techniques), only a few have targeted the medical
paradigm of the matter. Due to lack of medical facilities available in the remote areas, patients usually ignore
early symptoms which may worsen the situation as time progresses. Hence, there is a rising need for
automatic skin disease detection system with high accuracy. Thus, we develop a multiclass deep learning
model to differentiate between Healthy Skin Vs Skin suffering from a Disease and Classification of Skin
Diseases into its main class like Psoriasis with its subclasses Guttate, Plaque and Nail Psoriasis. We have used
Deep Learning to train our model, Deep Learning is a part of Machine Learning in which unlike Machine
Learning it uses large dataset and hence the number of classifiers is reduced substantially. The machine learns
itself and divide the data provided into the levels of prediction and in a very short period of time gives the
accurate results, thereby promoting and supporting development of Dermatology. The algorithm that we have
used is Convolutional Neural Network (CNN) as it is one of the most preferred algorithm for image
classification.

Keywords—Dermatoscopic images, Deep Learning, Data Enhancement, Convolutional Neural

Network(CNN), Model Training, Testing and Evaluation

I. INTRODUCTION

Skin diseases encompass a diverse range of conditions that affect the skin's health and appearance. From
common issues like acne and eczema to more complex and potentially severe conditions such as psoriasis,
fungal infections, and skin cancers, these ailments impact individuals across various age groups. Skin diseases
are often underestimated, with some perceiving them primarily as cosmetic concerns. However, their effects
extend beyond the surface, influencing overall health and quality of life. Symptoms like pain, itching,
discomfort, and emotional distress can significantly burden individuals. Timely diagnosis and management are
crucial to prevent complications, such as secondary infections and scarring. The prevalence of skin diseases
highlights the need for increased awareness, accessibility to dermatological care, and proactive approaches to
preventive care. By addressing these aspects, we can contribute to improved public health outcomes and
enhance the overall well-being of individuals affected by skin conditions Therefore, we propose an image
processing-based approach to diagnose the skin diseases. This method takes the digital image of disease effect
skin area then use image analysis to identify the type of disease. Our proposed approach is simple, fast and
does not require expensive equipment's other than a camera and a device like computer or mobile phone.

II. EXISTING TECHNOLOGY

Several researchers have proposed different image processing-based techniques to detect different type of
skin diseases. Most of the diagnosing methods rely on imaging technology, and the epidermal recognition of
such skin diseases does not need radiological imaging technologies. They can recognize the condition based
on the standard images through image processing techniques, including image transformation, equalization,
enhancement, edge detection, and segmentation. Here we briefly review some of the techniques as reported
in the references.
2.1. Artificial Neural Networks (ANN)
Artificial Neural Network (ANN) refers to a computational model inspired by the structure and functioning of
the human brain's neural networks. It is a fundamental component of machine learning and artificial
intelligence. Comprising interconnected nodes, or artificial neurons, organized into layers, ANNs are designed
to process information in a manner similar to the human brain. This model used [1] MobileNet V2 with the
LSTM and considered input image and the MobileNet V2 with LSTM model’s resultant outputs have no
significant randomness to explore all possible patterns in the assessment process. Alongside the bottleneck in
residual connections in the proposed architecture, the model yields higher accuracy with minimal effort.

II.2. Bayesian classification

The utilization of Bayesian classification is a common approach in skin disease classification [2] . This method
involves classifying images within various trained disease image datasets. However, the Naïve Bayes
classification encounters challenges with independent predictors, and the zero-probability problem makes its
implementation difficult in multi-objective-based domains. Naïve Bayes classifiers are not well-suited for
handling unsupervised data classification [3].

II.3. Decision tree

The Decision tree [4] algorithm is extensively employed for classifying skin diseases and predicting lower limb
ulcers and cervical cancer. However, this model requires extensive training and a significant level of accuracy.
Even a minor alteration in input data can lead to an exponential change in outcomes, rendering the model
insatiable. Furthermore, the Decision Tree model demands a relatively higher memory capacity, consequently
requiring more computational time [5].

II.4. K Nearest Neighbor

The K-Nearest Neighbor (KNN) model [6] is a widely employed classification method in forecasting and
predictive modeling. It is known for its simplicity and does not require extensive training. The KNN model
exhibits a notably high accuracy [7]. However, its suitability diminishes with larger-size datasets, as it may
significantly prolong the prediction process. Additionally, the model tends to underperform when handling
high-dimensional data with inappropriate feature information, which can impact its accuracy [8],
Consequently, KNN may not be the most suitable choice for skin disease classification.

Paper Link:
https://airl.csu.edu.cn/PDFs/LABELS2019_XiangyaDerm.pdf

II.5. Convolutional Neural Network

In this [], each image was confirmed by a disease label and was marked by a specialist with bounding boxes. It
is a large-scale, Asian-dominated dataset of skin diseases. It used 107,565 clinical images that cover 541 types
of skin diseases (i.e. 199 images for each type of skin disease) which identified that InceptionResNetV2 is the
best one for 80 skin disease classification for their dataset.

Paper Link:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6129338/

II.6. Support Vector Machine (SVM)

It used images to classify Herpes, Dermatitis, and Psoriasis. It preprocessed images to remove noise and
irrelevant background by filtering and transformations and used the method of grey-level co-occurrence
matrix (GLCM) to segment images of skin disease. It accurately obtained the texture and color features of
different skin disease images by using the Support Vector Machine (SVM) classification method.

III Dataset

III FUTURE WORK

To enhance and expand our application's capabilities, we aim to incorporate additional features to bolster its
robustness. Our future plans encompass the inclusion of more skin disease predictions, such as Eczema and
Shingles, including their subtypes. Furthermore, we intend to extend our detection capabilities to the Dermis
layer of the skin. The upcoming enhancements will not only predict diseases but also provide valuable insights
into precautions and cures for the identified conditions. Additionally, the application will offer information on
nearby clinics for consultation and treatment, complete with a map for easy navigation. This comprehensive
approach ensures a more holistic and user-friendly experience for our application's users.

REFERENCES

1. Mehdy, M.; Ng, P.; Shair, E.F.; Saleh, N.; Gomes, C. Artificial Neural Networks in Image Processing for
Early Detection of Breast Cancer. Comput. Math. Methods Med. 2017, 2017, 2610628.
2. Hsu, W. Bayesian Classification. Encyclopedia of Database Systems; Liu, L., Özsu, M.T., Eds.; Springer:
New York, NY, USA, 2018. [Google Scholar]
3. Tahmassebi, A.; Gandomi, A.; Schulte, M.; Goudriaan, A.; Foo, S.; Meyer-Base, A. Optimized Naive-
Bayes and Decision Tree Approaches for fMRI Smoking Cessation Classification. Complexity 2018, 2018,
1–24. [Google Scholar] [CrossRef]
4. Seixas, J.L.; Mantovani, R.G. Decision Trees for the Detection of Skin Lesion Patterns in Lower Limbs
Ulcers. In Proceedings of the 2016 International Conference on Computational Science and
Computational Intelligence (CSCI), Las Vegas, NV, USA, 15–17 December 2018; pp. 677–681. [ Google
Scholar]
5. Hamad, M.A.; Zeki, A.M. Accuracy vs. Cost in Decision Trees: A Survey. In Proceedings of the 2018
International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies
(3ICT), Sakhier, Bahrain, 18–20 November 2018; pp. 1–4. [Google Scholar]
6. Dang, Y.; Jiang, N.; Hu, H.; Ji, Z.; Zhang, W. Image classification based on quantum K-Nearest-Neighbor
algorithm. Quantum Inf. Process. 2018, 17, 1–18. [Google Scholar] [CrossRef]
7. Sumithra, R.; Suhil, M.; Guru, D.S. Segmentation and classification of skin lesions for disease
diagnosis. Procedia Comput. Sci. 2015, 45, 76–85.
8. Zhang, S.; Wu, Y.; Chang, J. Survey of Image Recognition Algorithms. In Proceedings of the 2020 IEEE
4th Information Technology, Networking, Electronic and Automation Control Conference (ITNEC),
Chongqing, China, 12–14 June 2020; pp. 542–548. [Google Scholar]
9. Dhandra B, Soma S, Reddy S, Mukarambi G. Color Histogram Approach for Analysis of Psoriasis Skin
Disease. InInt. Conf. on Multimedia Processing 2013.
10. Juang LH, Wu MN. Psoriasis image identification using k-means clustering with morphological
processing. Measurement. 44(5):895-905, 2011.
11. Pal A, Garain U, Chandra A, Chatterjee R, Senapati S. Psoriasis skin biopsy image segmentation
using Deep Convolutional Neural Network. Computer methods and programs in biomedicine. 159:59-
69, 2018.
12. Shrivastava VK, Londhe ND, Sonawane RS, Suri JS. Reliable and accurate psoriasis disease
classification in dermatology images using comprehensive feature space in machine learning
paradigm.
13. Shrivastava VK, Londhe ND, Sonawane RS, Suri JS. Computer-aided diagnosis of psoriasis skin images
with HOS, texture and color features: a first comparative study of its kind. Computer methods and
programs in biomedicine. 126:98-109,2016.
14. MohdAffandi A, Khan I, NgahSaaya N. Epidemiology and Clinical Features of Adult Patients with
Psoriasis in Malaysia: 10-Year Review from the Malaysian Psoriasis Registry (2007–2016). Dermatology
research and practice. 2018.
15. Withana U, Fernando P. Differential diagnosis of eczema and psoriasis using categorical data in image
processing. In 2017 Seventeenth International Conference on Advances in ICT for Emerging Regions
(ICTer) 2017 Sep 6 (pp. 1-6). IEEE.
16. Lu J, Kazmierczak E, Manton JH, Sinclair R. Automatic segmentation of scaling in 2-d psoriasis skin
images. IEEE transactions on medical imaging. 32(4):719-30, 2012.