Related Experiment Video
Updated: Aug 16, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Deep Learning Approach in Image Diagnosis of Pseudomonas Keratitis
Ming-Tse Kuo1,2, Benny Wei-Yun Hsu3, Yi Sheng Lin3
1Department of Ophthalmology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung 83301, Taiwan.
Abstract:
This investigation aimed to explore deep learning (DL) models' potential for diagnosing Pseudomonas keratitis using external eye images. In the retrospective research, the images of bacterial keratitis (BK, n = 929), classified as Pseudomonas (n = 618) and non-Pseudomonas (n = 311) keratitis, were collected. Eight DL algorithms, including ResNet50, DenseNet121, ResNeXt50, SE-ResNet50, and EfficientNets B0 to B3, were adopted as backbone models to train and obtain the best ensemble 2-, 3-, 4-, and 5-DL models. Five-fold cross-validation was used to determine the ability of single and ensemble models to diagnose Pseudomonas keratitis. The EfficientNet B2 model had the highest accuracy (71.2%) of the eight single-DL models, while the best ensemble 4-DL model showed the highest accuracy (72.1%) among the ensemble models. However, no statistical difference was shown in the area under the receiver operating characteristic curve and diagnostic accuracy among these single-DL models and among the four best ensemble models. As a proof of concept, the DL approach, via external eye photos, could assist in identifying Pseudomonas keratitis from BK patients. All the best ensemble models can enhance the performance of constituent DL models in diagnosing Pseudomonas keratitis, but the enhancement effect appears to be limited.
Insights
Deep learning models show potential for diagnosing Pseudomonas keratitis from eye images. While ensemble models slightly improved accuracy, the enhancement effect was limited, suggesting further research is needed for clinical application.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Bacterial keratitis (BK) is a serious eye infection.
- Accurate diagnosis of *Pseudomonas* keratitis is crucial for effective treatment.
- External eye images offer a non-invasive diagnostic approach.
Purpose of the Study:
- To evaluate the efficacy of deep learning (DL) models in diagnosing *Pseudomonas* keratitis using external eye images.
- To compare the performance of various single and ensemble DL models.
- To assess the potential of DL as a diagnostic aid for *Pseudomonas* keratitis.
Main Methods:
- Retrospective analysis of 929 bacterial keratitis (BK) images, including 618 *Pseudomonas* and 311 non-*Pseudomonas* cases.
- Training and evaluation of eight DL algorithms (ResNet50, DenseNet121, ResNeXt50, SE-ResNet50, EfficientNets B0-B3) as single and ensemble models.
- Utilized five-fold cross-validation to assess diagnostic capabilities.
Main Results:
- The EfficientNet B2 model achieved the highest accuracy (71.2%) among single DL models.
- The best performing ensemble model (4-DL) showed an accuracy of 72.1%.
- No statistically significant differences were observed in diagnostic accuracy or AUC among the evaluated single and ensemble models.
Conclusions:
- Deep learning models, applied to external eye photographs, can potentially assist in identifying *Pseudomonas* keratitis.
- Ensemble DL models demonstrated a limited enhancement in diagnostic performance compared to individual models.
- Further development is needed to optimize DL models for improved clinical utility in *Pseudomonas* keratitis diagnosis.
