-
In this project, I used to extract the face feature (Eye, Nose, Eyebrow, and Mouth) using the DLib library and pass each feature to each of the multi-stream CNN layers. After passing Through Multi-stream CNN the output is concatenated and again passes through the CNN model, the output of CNN is estimated age.
-
For a more detailed technical report, check here
- We trained our model with a downsampled version of Adience dataset