| Technical Name |
人工智慧於息肉狀脈絡膜血管病變之診斷與治療決策之應用 |
| Project Operator |
Taipei Veterans General Hospital |
| Project Host |
1 劉瑞玲 2 盧鴻興 (總主持人：張德明) |
| Summary |
This system help doctors to distinguish age-related macular degeneration (ARMD)polypoidal choroidal vasculopathy (PCV) by color fundus photography. We have collected large image dataset from Taipei Veterans General Hospitalother hospitals to establishvalidate CNN model, which had been tested by external image datasetmodified to improve the accuracy of the AI model. |
| Scientific Breakthrough |
In usual clinical practice by ophthalmologist, it is time-consuminginvasive to differentiate PCV from ARMD by the golden-standard examination. However, using this deep learning model only by entering color fundus photography can help us to rapidly distinguish ARMDPCV, the accuracy can reach up to 94.52 (EfficientNet). The GradCAM picture confirms the precise location of the lesions. |
| Industrial Applicability |
This technique can help doctors to distinguish ARMDPCV, especially in rural area where is lack of medical resources. It can also help us to prevent lengthyinvasive medical examination. Besides, it is more useful in the telemedicine. In the future, if we can combine this technique to portable devicecloud-based platform, it will be more powerful to improve the quality of medication. |
| Keyword |
Age-related macular degeneration Polypoidal Choroidal Vasculopathy Deep learning Artificial intelligence Ophthalmology EfficientNet GradCAM |
