An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms
The identification of Chinese herbal medicines (CHMs) is directly related to their safety and efficacy. However, the limited number of professionals fails to meet the demand of the vast CHMs market. To make the CHMs identification more convenient and accurate, this study aimed at assessing the feasi...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2021-01-01
|
| Series: | Informatics in Medicine Unlocked |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2352914821000976 |
| _version_ | 1818738003847151616 |
|---|---|
| author | Wenting Chen Jiayun Tong Rui He Ying Lin Peiting Chen Zixin Chen Xiaoli Liu |
| author_facet | Wenting Chen Jiayun Tong Rui He Ying Lin Peiting Chen Zixin Chen Xiaoli Liu |
| author_sort | Wenting Chen |
| collection | DOAJ |
| description | The identification of Chinese herbal medicines (CHMs) is directly related to their safety and efficacy. However, the limited number of professionals fails to meet the demand of the vast CHMs market. To make the CHMs identification more convenient and accurate, this study aimed at assessing the feasibility of the automated machine learning (AutoML) technology in CHMs image recognition. Instead of a handcrafted neural network experimental models built on AutoML platform were presented. A dataset of 31,460 images consisting of 315 categories of commonly-used CHMs was built for the model creation. Furthermore, the Huawei ModelArts model was compared with a model built on the Baidu EasyDL platform using the same dataset. Three professionals were also invited to recognize images of 315 categories of CHMs. During the model evaluation, high accuracies of 99.2% and 98.4% were achieved by ModelArts and EasyDL, respectively. In the held-out tests, the accuracies of ModelArts and EasyDL models were 91.2% and 91.85%, respectively. Both models performed very well individually and no statistically significant difference was found in model performance between them. However, the model-training time was only approximately 41 min on ModelArts platform but 118 min on EasyDL. The mean accuracy of the manual recognition for 315 CHMs was 97.46 ± 1.58%. Results revealed that AutoML technology is a fast and simple approach and has great practical potential in the field of CHMs image recognition. Since the Huawei ModelArts platform requires less training time, we recommend it as a priority. |
| first_indexed | 2024-12-18T01:02:02Z |
| format | Article |
| id | doaj.art-5c951f3c54e24382a5dc032727e8343a |
| institution | Directory Open Access Journal |
| issn | 2352-9148 |
| language | English |
| last_indexed | 2024-12-18T01:02:02Z |
| publishDate | 2021-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Informatics in Medicine Unlocked |
| spelling | doaj.art-5c951f3c54e24382a5dc032727e8343a2022-12-21T21:26:21ZengElsevierInformatics in Medicine Unlocked2352-91482021-01-0125100607An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platformsWenting Chen0Jiayun Tong1Rui He2Ying Lin3Peiting Chen4Zixin Chen5Xiaoli Liu6School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaCorresponding author.; School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaSchool of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaSchool of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaSchool of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaSchool of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaSchool of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, 510006, People's Republic of ChinaThe identification of Chinese herbal medicines (CHMs) is directly related to their safety and efficacy. However, the limited number of professionals fails to meet the demand of the vast CHMs market. To make the CHMs identification more convenient and accurate, this study aimed at assessing the feasibility of the automated machine learning (AutoML) technology in CHMs image recognition. Instead of a handcrafted neural network experimental models built on AutoML platform were presented. A dataset of 31,460 images consisting of 315 categories of commonly-used CHMs was built for the model creation. Furthermore, the Huawei ModelArts model was compared with a model built on the Baidu EasyDL platform using the same dataset. Three professionals were also invited to recognize images of 315 categories of CHMs. During the model evaluation, high accuracies of 99.2% and 98.4% were achieved by ModelArts and EasyDL, respectively. In the held-out tests, the accuracies of ModelArts and EasyDL models were 91.2% and 91.85%, respectively. Both models performed very well individually and no statistically significant difference was found in model performance between them. However, the model-training time was only approximately 41 min on ModelArts platform but 118 min on EasyDL. The mean accuracy of the manual recognition for 315 CHMs was 97.46 ± 1.58%. Results revealed that AutoML technology is a fast and simple approach and has great practical potential in the field of CHMs image recognition. Since the Huawei ModelArts platform requires less training time, we recommend it as a priority.http://www.sciencedirect.com/science/article/pii/S2352914821000976Chinese herbal medicines (CHMs)IdentificationAutomated machine learning (AutoML)Image recognitionHuawei ModelArtsBaidu EasyDL |
| spellingShingle | Wenting Chen Jiayun Tong Rui He Ying Lin Peiting Chen Zixin Chen Xiaoli Liu An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms Informatics in Medicine Unlocked Chinese herbal medicines (CHMs) Identification Automated machine learning (AutoML) Image recognition Huawei ModelArts Baidu EasyDL |
| title | An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms |
| title_full | An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms |
| title_fullStr | An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms |
| title_full_unstemmed | An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms |
| title_short | An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms |
| title_sort | easy method for identifying 315 categories of commonly used chinese herbal medicines based on automated image recognition using automl platforms |
| topic | Chinese herbal medicines (CHMs) Identification Automated machine learning (AutoML) Image recognition Huawei ModelArts Baidu EasyDL |
| url | http://www.sciencedirect.com/science/article/pii/S2352914821000976 |
| work_keys_str_mv | AT wentingchen aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT jiayuntong aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT ruihe aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT yinglin aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT peitingchen aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT zixinchen aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT xiaoliliu aneasymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT wentingchen easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT jiayuntong easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT ruihe easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT yinglin easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT peitingchen easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT zixinchen easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms AT xiaoliliu easymethodforidentifying315categoriesofcommonlyusedchineseherbalmedicinesbasedonautomatedimagerecognitionusingautomlplatforms |