FlatFeet classification

FlatFeet x-ray image classification project

1. Input Image Preprocessing

2. Train

• train image : 560
• valid image : 80
• class : 2 (normal / flatfeet)
• optimizer : Adam
• epoch : 300 (Early stopping)
• learning rate : 0.001
• batch size : 8
• model : M5(custom dual input model)

3. Test Acc & Confusion Matrix

Test Image : 80 (normal 40 + flatfeet 40)

• test accuracy : 71.25%
• confusion matrix