This code can detect age categories (YOUNG, MIDDLE, OLD) for a given image
This can be used in mobile devices, scanners to detect ages, can be a fun app.
The example code is in Python (version 2.7 or higher will work).
.import cv2
.import pandas, numpy
.import keras
.import sklearn
A computer vision system that can automatically detect age categories in an image, that is trained on 19k+ images.
Each image is standardized with size 32x32x3 and trained on CNN coupled with a feed forward neural network for classification.
Run the predictions.ipynb code by following the steps as mentioned in the code.
Accuracy = 78.1%
- Changing the architecture and hyper parameters to finetune the training would improve accuracies.
- Images can be denoised and then can be given to training. like, Concolutional AutoEncoders
- Adding further preprocessing techniques/filters can improve the quality of training.