Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification
Recognizing insect pests using images is an important and challenging research issue. A correct species classification will help choosing a more proper mitigation strategy regarding crop management, but designing an automated solution is also difficult due to the high similarity between species at s...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Agronomy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4395/12/8/1733 |
| _version_ | 1797411667415400448 |
|---|---|
| author | Jacó C. Gomes Díbio L. Borges |
| author_facet | Jacó C. Gomes Díbio L. Borges |
| author_sort | Jacó C. Gomes |
| collection | DOAJ |
| description | Recognizing insect pests using images is an important and challenging research issue. A correct species classification will help choosing a more proper mitigation strategy regarding crop management, but designing an automated solution is also difficult due to the high similarity between species at similar maturity stages. This research proposes a solution to this problem using a few-shot learning approach. First, a novel insect data set based on curated images from IP102 is presented. The IP-FSL data set is composed of 97 classes of adult insect images, and 45 classes of early stages, totalling 6817 images. Second, a few-shot prototypical network is proposed based on a comparison with other state-of-art models and further divergence analysis. Experiments were conducted separating the adult classes and the early stages into different groups. The best results achieved an accuracy of 86.33% for the adults, and 87.91% for early stages, both using a Kullback–Leibler divergence measure. These results are promising regarding a crop scenario where the more significant pests are few and it is important to detect them at earlier stages. Further research directions would be in evaluating a similar approach in particular crop ecosystems, and testing cross-domains. |
| first_indexed | 2024-03-09T04:49:27Z |
| format | Article |
| id | doaj.art-049b7ef727e948cd8e4073f2919ffb58 |
| institution | Directory Open Access Journal |
| issn | 2073-4395 |
| language | English |
| last_indexed | 2024-03-09T04:49:27Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Agronomy |
| spelling | doaj.art-049b7ef727e948cd8e4073f2919ffb582023-12-03T13:11:18ZengMDPI AGAgronomy2073-43952022-07-01128173310.3390/agronomy12081733Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages ClassificationJacó C. Gomes0Díbio L. Borges1Department of Mechanical Engineering, University of Brasília, Brasília 70910-900, BrazilDepartment of Computer Science, University of Brasília, Brasília 70910-900, BrazilRecognizing insect pests using images is an important and challenging research issue. A correct species classification will help choosing a more proper mitigation strategy regarding crop management, but designing an automated solution is also difficult due to the high similarity between species at similar maturity stages. This research proposes a solution to this problem using a few-shot learning approach. First, a novel insect data set based on curated images from IP102 is presented. The IP-FSL data set is composed of 97 classes of adult insect images, and 45 classes of early stages, totalling 6817 images. Second, a few-shot prototypical network is proposed based on a comparison with other state-of-art models and further divergence analysis. Experiments were conducted separating the adult classes and the early stages into different groups. The best results achieved an accuracy of 86.33% for the adults, and 87.91% for early stages, both using a Kullback–Leibler divergence measure. These results are promising regarding a crop scenario where the more significant pests are few and it is important to detect them at earlier stages. Further research directions would be in evaluating a similar approach in particular crop ecosystems, and testing cross-domains.https://www.mdpi.com/2073-4395/12/8/1733few-shot learninginsect pest classificationinsect maturity stagesRGB images |
| spellingShingle | Jacó C. Gomes Díbio L. Borges Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification Agronomy few-shot learning insect pest classification insect maturity stages RGB images |
| title | Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification |
| title_full | Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification |
| title_fullStr | Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification |
| title_full_unstemmed | Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification |
| title_short | Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification |
| title_sort | insect pest image recognition a few shot machine learning approach including maturity stages classification |
| topic | few-shot learning insect pest classification insect maturity stages RGB images |
| url | https://www.mdpi.com/2073-4395/12/8/1733 |
| work_keys_str_mv | AT jacocgomes insectpestimagerecognitionafewshotmachinelearningapproachincludingmaturitystagesclassification AT dibiolborges insectpestimagerecognitionafewshotmachinelearningapproachincludingmaturitystagesclassification |