-
Image size
-
precision, recall, f1-score, support
precision recall f1-score support others 0.92 0.78 0.84 59 pavbhaji 0.84 0.94 0.89 70 micro avg 0.87 0.87 0.87 129 macro avg 0.88 0.86 0.86 129 weighted avg 0.87 0.87 0.87 129 confusion matrix [[46 13] [ 4 66]]
Model: xception This particular model helps to reduce model parameters with depthwise separable convolution.
- Data set is very small to segregate into categories.
- Image has noisy background.
- Imbalanced data set
I used augmentation technique and pre-trained imageNet weights.