Mineral Identification Based on Multi-Label Image Classification
The identification of minerals is indispensable in geological analysis. Traditional mineral identification methods are highly dependent on professional knowledge and specialized equipment which often consume a lot of labor. To solve this problem, some researchers use machine learning algorithms to q...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-10-01
|
| Series: | Minerals |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-163X/12/11/1338 |
| _version_ | 1797467182516404224 |
|---|---|
| author | Baokun Wu Xiaohui Ji Mingyue He Mei Yang Zhaochong Zhang Yan Chen Yuzhu Wang Xinqi Zheng |
| author_facet | Baokun Wu Xiaohui Ji Mingyue He Mei Yang Zhaochong Zhang Yan Chen Yuzhu Wang Xinqi Zheng |
| author_sort | Baokun Wu |
| collection | DOAJ |
| description | The identification of minerals is indispensable in geological analysis. Traditional mineral identification methods are highly dependent on professional knowledge and specialized equipment which often consume a lot of labor. To solve this problem, some researchers use machine learning algorithms to quickly identify a single mineral in images. However, in the natural environment, minerals often exist in an associated form, which makes the identification impossible with traditional machine learning algorithms. For the identification of associated minerals, this paper proposes a deep learning model based on the transformer and multi-label image classification. The model uses transformer architecture to model mineral images and outputs the probability of the existence of various minerals in an image. The experiments on 36 common minerals show that the model can achieve a mean average precision of 85.26%. The visualization of the class activation mapping indicates that our model can roughly locate the identified minerals. |
| first_indexed | 2024-03-09T18:49:02Z |
| format | Article |
| id | doaj.art-7cb7b16803f4409f909aa02ba9417556 |
| institution | Directory Open Access Journal |
| issn | 2075-163X |
| language | English |
| last_indexed | 2024-03-09T18:49:02Z |
| publishDate | 2022-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Minerals |
| spelling | doaj.art-7cb7b16803f4409f909aa02ba94175562023-11-24T05:57:50ZengMDPI AGMinerals2075-163X2022-10-011211133810.3390/min12111338Mineral Identification Based on Multi-Label Image ClassificationBaokun Wu0Xiaohui Ji1Mingyue He2Mei Yang3Zhaochong Zhang4Yan Chen5Yuzhu Wang6Xinqi Zheng7School of Information Engineering, China University of Geosciences, Beijing 100083, ChinaSchool of Information Engineering, China University of Geosciences, Beijing 100083, ChinaGemological Institute, China University of Geosciences, Beijing 100083, ChinaSciences Institute, China University of Geosciences, Beijing 100083, ChinaSchool of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, ChinaState Key Laboratory of Biogeology and Environmental Geology, School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, ChinaSchool of Information Engineering, China University of Geosciences, Beijing 100083, ChinaSchool of Information Engineering, China University of Geosciences, Beijing 100083, ChinaThe identification of minerals is indispensable in geological analysis. Traditional mineral identification methods are highly dependent on professional knowledge and specialized equipment which often consume a lot of labor. To solve this problem, some researchers use machine learning algorithms to quickly identify a single mineral in images. However, in the natural environment, minerals often exist in an associated form, which makes the identification impossible with traditional machine learning algorithms. For the identification of associated minerals, this paper proposes a deep learning model based on the transformer and multi-label image classification. The model uses transformer architecture to model mineral images and outputs the probability of the existence of various minerals in an image. The experiments on 36 common minerals show that the model can achieve a mean average precision of 85.26%. The visualization of the class activation mapping indicates that our model can roughly locate the identified minerals.https://www.mdpi.com/2075-163X/12/11/1338mineral identificationdeep learningtransformerconvolutional neural networkmulti-label image classification |
| spellingShingle | Baokun Wu Xiaohui Ji Mingyue He Mei Yang Zhaochong Zhang Yan Chen Yuzhu Wang Xinqi Zheng Mineral Identification Based on Multi-Label Image Classification Minerals mineral identification deep learning transformer convolutional neural network multi-label image classification |
| title | Mineral Identification Based on Multi-Label Image Classification |
| title_full | Mineral Identification Based on Multi-Label Image Classification |
| title_fullStr | Mineral Identification Based on Multi-Label Image Classification |
| title_full_unstemmed | Mineral Identification Based on Multi-Label Image Classification |
| title_short | Mineral Identification Based on Multi-Label Image Classification |
| title_sort | mineral identification based on multi label image classification |
| topic | mineral identification deep learning transformer convolutional neural network multi-label image classification |
| url | https://www.mdpi.com/2075-163X/12/11/1338 |
| work_keys_str_mv | AT baokunwu mineralidentificationbasedonmultilabelimageclassification AT xiaohuiji mineralidentificationbasedonmultilabelimageclassification AT mingyuehe mineralidentificationbasedonmultilabelimageclassification AT meiyang mineralidentificationbasedonmultilabelimageclassification AT zhaochongzhang mineralidentificationbasedonmultilabelimageclassification AT yanchen mineralidentificationbasedonmultilabelimageclassification AT yuzhuwang mineralidentificationbasedonmultilabelimageclassification AT xinqizheng mineralidentificationbasedonmultilabelimageclassification |