음각 정보를 이용한 딥러닝 기반의 알약 식별 알고리즘 연구 = Pill Identification Algorithm Based on Deep Learning Using Imprinted Text Feature

다국어 초록 (Multilingual Abstract)

In this paper, we propose a pill identification model using engraved text feature and image feature such as shape and color, and compare it with an identification model that does not use engraved text feature to verify the possibility of improving identification performance by improving recognition rate of the engraved text. The data con- sisted of 100 classes and used 10 images per class. The engraved text feature was acquired through Keras OCR based on deep learning and 1D CNN, and the image feature was acquired through 2D CNN. According to the identification results, the accuracy of the text recognition model was 90%. The accuracy of the comparative model and the proposed model was 91.9% and 97.6%. The accuracy, precision, recall, and F1-score of the proposed model were better than those of the comparative model in terms of statistical significance. As a result, we confirmed that the expansion of the range of feature improved the performance of the identification model.

참고문헌 (Reference)

1 김영재 ; 김광기, "의료 영상에 최적화된 딥러닝 모델의 개발" 대한영상의학회 81 (81): 1274-1289, 2020

2 엄고혜 ; 이주연 ; 조윤희 ; 조윤숙 ; 한현주 ; 손인자, "약사의 중재활동에 의한 약반납 감소효과에 관한 연구" 한국병원약사회 23 (23): 1-9, 2006

3 이경윤 ; 김영재 ; 김승태 ; 김효은 ; 김광기, "알약 자동 인식을 위한 딥러닝 모델간 비교 및 검증" 한국멀티미디어학회 22 (22): 349-356, 2019

4 오다은 ; 이미수 ; 이윤덕 ; 예경남 ; 김정태, "미반납약 감소를 위한 약품 반납 업무 개선 활동" 한국병원약사회 28 (28): 364-371, 2011

5 Albawi S, "Understanding of a convolutional neural network" 1-6, 2017

6 Vijayarani S, "Text mining: open source tokenization tools-an analysis" 3 (3): 37-47, 2016

7 Kim DW, "Shape and Text Imprint Recognition of Pill Image Taken with a Smartphone" Seoul National University 2017

8 Kudo T, "Sentencepiece : A simple and language independent subword tokenizer and detokenizer for neural text processing" 66-71, 2018

9 Du C, "Selective feature connection mechanism : Concatenating multi-layer CNN fea-tures with a feature selector" 129 : 108-114, 2020

10 Lei Z, "Scene text recognition using residual convolutional recurrent neural network" 29 (29): 861-871, 2018

11 Yacouby R, "Probabilistic extension of precision, recall, and F1 score for more thorough evaluation of classi-fication models" 79-79, 2020

12 Wang Y, "Pill Recogni-tion Using Minimal Labeled Data" 346-353, 2017

13 Ko DG., "Optical Character Recognition Performance Com-parison of CNNs and Tesseract" Sungkyunk-wan University 2016

14 Chotivatunyu P, "Medicine Identification System on Mobile Devices for the Elderly" 1-6, 2020

15 Firoz R, "Medical image enhancement using mor-phological transformation" 4 (4): 1-12, 2016

16 Sun M, "Learning pooling for convolutional neural network" 224 : 96-104, 2017

17 Maitrichit N, "Intelligent Medicine Identifica-tion System Using a Combination of Image Recognition and Optical Character Recognition" 1-5, 2020

18 Larios Delgado N, "Fast and accurate medication identification" 2 (2): 1-9, 2019

19 Dwarampudi M, "Effects of padding on LSTMs and CNNs"

20 Erkan U, "Different applied median filter in salt and pepper noise" 70 : 789-798, 2018

21 Gulli A, "Deep learning with TensorFlow 2and Keras: regression, ConvNets, GANs, RNNs, NLP, and more with TensorFlow 2 and the Keras API" Packt Publishing Ltd 63-66, 2019

22 Chollet F, "Deep learning with Python" Simon and Schuster 62-64, 2021

23 Manaswi NK, "Deep Learning with Applications Using Python" Apress 31-43, 2018

24 Ertam F, "Data classification with deep learning using Tensorflow" 755-758, 2017

25 Baek Y, "Character region aware-ness for text detection" 9365-9374, 2019

26 Xu Z, "Canny edge detection based on Open CV" 53-56, 2017

27 Suntronsuk S, "Automatic text imprint analysis from pill images" 288-293, 2017

28 Ou YY, "Automatic drug pills detection based on enhanced feature pyramid network and convolution neural networks" 14 (14): 9-17, 2020

29 Sahu S, "An approach for de-noising and contrast enhancement of retinal fundus image using CLAHE" 110 : 87-98, 2019

30 Hassanein A. S, "A survey on Hough transform, theory, techniques and applications" 12 (12): 139-156, 2015

31 Chicho BT, "A Comprehensive Survey of Deep Learning Models Based on Keras Framework" 2 (2): 49-62, 2021