Back

Addressing domain shift in deep learning: Challenges and insights from plant disease diagnosis and flower recognition

Sun, J.

2024-10-11 plant biology
10.1101/2024.10.07.617111 bioRxiv
Show abstract

Deep learning models have shown significant potential for plant pest and disease (PPD) diagnosis; however, their real-world effectiveness is often limited by variability between datasets, where models trained on one dataset perform poorly on others collected under different conditions. In this study, I evaluated the cross-dataset generalization of widely used deep learning architectures, including ResNet, EfficientNet, Inception, and MobileNet, across multiple tomato pest and disease datasets. As expected, models trained and tested on the same dataset achieved high performance. However, substantial performance degradation occurred when these models were tested on different datasets, highlighting the challenges posed by dataset variability. This trend was consistent across all evaluated architectures, indicating that changing the model architecture alone is insufficient to address these issues. The findings emphasize the need for more diverse and representative datasets to better capture variability in agricultural data and enhance the practical deployment of deep learning models for PPD diagnosis.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
Plant Phenomics
18 papers in training set
Top 0.1%
23.0%
2
The Plant Phenome Journal
14 papers in training set
Top 0.1%
10.0%
3
Frontiers in Plant Science
256 papers in training set
Top 0.8%
9.2%
4
New Phytologist
346 papers in training set
Top 1%
8.1%
50% of probability mass above
5
PLOS ONE
5266 papers in training set
Top 27%
5.7%
6
Scientific Data
209 papers in training set
Top 0.5%
5.6%
7
Phytopathology®
31 papers in training set
Top 0.2%
3.3%
8
PLANTS, PEOPLE, PLANET
27 papers in training set
Top 0.3%
2.2%
9
Scientific Reports
3612 papers in training set
Top 46%
2.2%
10
Plant Physiology
238 papers in training set
Top 2%
2.2%
11
Applications in Plant Sciences
23 papers in training set
Top 0.2%
1.8%
12
Plant Methods
42 papers in training set
Top 0.4%
1.8%
13
Biology Methods and Protocols
61 papers in training set
Top 1%
1.5%
14
GigaScience
212 papers in training set
Top 3%
1.5%
15
Quantitative Plant Biology
15 papers in training set
Top 0.1%
1.4%
16
PLOS Computational Biology
1863 papers in training set
Top 16%
1.2%
17
Plant Direct
95 papers in training set
Top 2%
1.2%
18
Remote Sensing in Ecology and Conservation
14 papers in training set
Top 0.2%
1.1%
19
in silico Plants
27 papers in training set
Top 0.2%
0.9%
20
Theoretical and Applied Genetics
49 papers in training set
Top 0.7%
0.9%
21
Molecular Plant-Microbe Interactions®
57 papers in training set
Top 0.9%
0.9%
22
G3: Genes, Genomes, Genetics
252 papers in training set
Top 5%
0.6%