Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Classification Model of Facial Acne Using Deep Learning

딥 러닝을 이용한 안면 여드름 분류 모델

  • Jung, Cheeoh (Department of Computer Engineering, Paichai University) ;
  • Yeo, Ilyeon (Department of Computer Engineering, Paichai University) ;
  • Jung, Hoekyung (Department of Computer Engineering, Paichai University)
  • Received : 2019.01.22
  • Accepted : 2019.04.04
  • Published : 2019.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

The limitations of applying a variety of artificial intelligence to the medical community are, first, subjective views, extensive interpreters and physical fatigue in interpreting the image of an interpreter's illness. And there are questions about how long it takes to collect annotated data sets for each illness and whether to get sufficient training data without compromising the performance of the developed deep learning algorithm. In this paper, when collecting basic images based on acne data sets, the selection criteria and collection procedures are described, and a model is proposed to classify data into small loss rates (5.46%) and high accuracy (96.26%) in the sequential structure. The performance of the proposed model is compared and verified through a comparative experiment with the model provided by Keras. Similar phenomena are expected to be applied to the field of medical and skin care by applying them to the acne classification model proposed in this paper in the future.

의학계에 다양하게 인공지능을 적용하는데 있어 한계는 우선적으로 해석자의 병증 이미지를 해석하는데 주관적 견해와 광범위한 해석자, 육체적 피로감 등이다. 그리고 병증마다 주석 달린 데이터 셋을 수집하는데 기간이 오래 걸린다는 것과 개발된 딥러닝 학습 알고리즘의 성능 저하가 없으면서도 충분한 훈련 데이터를 얻을지에 대한 의문이 있다는 것이다. 이에 본 논문에서는 여드름 데이터 셋을 기준으로 기본 이미지를 수집할 때 선정 기준과 수집 절차에 대해 연구하고, Sequential 구조로 딥 러닝 기법을 적용하여 적은 손실률(5.46%)과 높은 정확도(96.26%)로 데이터를 분류하는 모델을 제안한다. Keras에서 기본 제공하는 모델과 비교실험을 통해 제안 모델의 성능을 비교 검증한다. 향후 본 논문에서 제안하는 여드름 분류 모델에 유사 현상들 적용하여 의학 및 피부 관리 분야에도 적용 가능할 것으로 예상된다.

Keywords

HOJBC0_2019_v23n4_381_f0001.png 이미지

Fig. 1 Image Collection Procedure

HOJBC0_2019_v23n4_381_f0002.png 이미지

Fig. 2 Correlation between data amount and performance

HOJBC0_2019_v23n4_381_f0003.png 이미지

Fig. 3 Preprocessing of Dataset

HOJBC0_2019_v23n4_381_f0004.png 이미지

Fig. 4 Proposed Model Architecture

HOJBC0_2019_v23n4_381_f0005.png 이미지

Fig. 5 Overall structure and data flow of the proposed model

HOJBC0_2019_v23n4_381_f0006.png 이미지

Fig. 6 Loading prepared dataset

HOJBC0_2019_v23n4_381_f0007.png 이미지

Fig. 7 Network Definition

HOJBC0_2019_v23n4_381_f0008.png 이미지

Fig. 8 Loss rate and Accuracy Convergence Curve (Typical CNN model provided by Keras)

HOJBC0_2019_v23n4_381_f0009.png 이미지

Fig. 9 Loss rate and Accuracy Convergence Curve (VGG model provided by Keras)

Table. 1 Collected Images by Type

HOJBC0_2019_v23n4_381_t0001.png 이미지

Table. 2 Development Environment

HOJBC0_2019_v23n4_381_t0002.png 이미지

Table. 3 Conditions

HOJBC0_2019_v23n4_381_t0003.png 이미지

Table. 4 Result Composite Table

HOJBC0_2019_v23n4_381_t0004.png 이미지

References

  1. L. Liu, W. Ouyang, X. Wang, P. Fieguth, J. Chen, X. Liu, and M. Pietikainen, "Deep Learning for Generics Object Detection : A Survey", International Journal of Computer Vision, pp. 1-30, Sep. 2018.
  2. H. Pratt, F. Coenen, S. Harding, and Y. Zheng, "Convolutional Neural Networks for Diabetic Retinopathy," International Conference On Medical Imaging Understanding and Analysis, vol. 90, pp. 200-205, Jul. 2016.
  3. H. Bilen, and A. Vedaldi, "Weakly Supervised Deep Detection Networks," The IEEE Conference on Computer Vision and Pattern Recognition, pp. 2846-2854, Dec. 2016.
  4. C. Cao, F. Liu, H. Tan, D. Song, W. Shu, W. Li, Y. Zhou, X. Bo, and Z. Xie, "Deep Learning and Its Applications in Biomdicine," Genomics, Protenomics & Bioinformatics, vol. 16, Issue1, pp. 17-32, Feb. 2018. https://doi.org/10.1016/j.gpb.2017.07.003
  5. M. I. Razzak, S. Naz, and A. Zaib, "Deep Learning for Medical Image Processing: Overview, Challenges and Future," Computer Vision and Pattern Recognition, vol. 26, pp. 323-350, Nov. 2017.
  6. E. Klang, "Deep learning and medical imaging," Journal of Thoracic Disease, vol. 10, no. 3, pp. 1325-1328, Feb. 2018. https://doi.org/10.21037/jtd.2018.02.76
  7. G. Litjens, T. Kooi, B. E. Bejnordi, A. Arindra, M. Ghafoorian, and C. I.Sanchez, "A survey on deep learning in medical image analysis," Medical Image Analysis, vol. 42, pp. 60-88, Dec. 2017. https://doi.org/10.1016/j.media.2017.07.005
  8. A. L. Zaenglein, A. L. Pathy, B. J. Schlosser, A. Alikhan, H. E. Baldwin, D. S. Berson, W. P. Bowe, E. M. Graber, S. W. Kang, J. E. Keri, J. J. Leyden, R. Reynolds, N. B. SilverBerg, L. F. SteinGold, M. M. Tollefson, J. S. Weiss, N. C. Dolan, A. A. Sagan, M. Stern, K. M. Boyer, and R. Bhushan, "Guidelines of care for the management of acne vulgaris," Journal of the American Academy of Dermatology, vol. 74, no. 5, pp. 945-973, May. 2016. https://doi.org/10.1016/j.jaad.2015.12.037
  9. J. K. Than, L. F. Stein Gold, A. F. Alexis, and J. C. Harper, "Current Concepts in Acne Pathogenesis: Pathways to Inflammation," Seminars in Cutaneous Medicine and Surgery, vol. 37, no. 3S, pp. 60-62, Jun. 2018. https://doi.org/10.12788/J.SDER.2018.024

Cited by

  1. CNN 기반 독성 식물 판별 시스템 vol.24, pp.8, 2019, https://doi.org/10.6109/jkiice.2020.24.8.993
  2. Analysis of Factors Affecting Satisfaction of Korean Cosmetics by Chinese Consumers Using Statistical Analysis Techniques vol.25, pp.1, 2019, https://doi.org/10.6109/jkiice.2021.25.1.152