| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
| Name | Name | Last commit date | ||
|---|---|---|---|---|
I have created a face recognizer using convolutional neural networks to recognize the authorized person's faces. I have trained the model on 50 different images of 2 persons using google colab. The accuracy is more than 85%
The CNN model architecture is as follows:
Conv: 16 filters, 3x3 kernel; MaxPooling: 2x2
Conv: 32 filters, 3x3 kernel; MaxPooling: 2x2
Conv: 64 filters, 3x3 kernel; MaxPooling: 2x2
Dense: 512 nodes
Dense: 1 node (output)
| Back | FazBrowse Home | New Git URL |