/Age_Estimation

Age Estimation

Primary LanguageJupyter NotebookMIT LicenseMIT

Age Estimation

This task solved with this repo

  • Age Estimation: I trained a regression model on UTKFace Dataset to estimate a person's age based on human facial appearance, using ResNet50 convolutional neural network

How to install

Run this command:

pip install -r requirements.txt

You can also download my age estimator model from here

Traning Information:

  • Validation Loss: 5.6115
  • Training Loss: 2.4291

How to process on an image:

Note: I used OpenCV cascading to detect faces. You can use any other face detector.

Run this command

python3 processing_on_image.py --img YOUR_IMAGE.JPG

You can also see the other arguments of it with this command

python processing_on_image.py --help

For Example:

  • --model: The age estimator model file path default:weights/best_age_estimator_resnet.h5
  • --draw, --no-draw: You can draw bounding boxes on the image. default:True
  • --show, --no-show: You can see the result by this command default:True
  • --save, --no-save: You can save the result by this command default:True

Result of age estimator on Messi image:

Result Of Identification

How to process on a video:

Run this command

python3 main.py --source YOUR_VIDEO_PATH OR CAMERA_ID

You can also see the other arguments of it with this command

python main.py --help

For Example:

  • --model: The age estimator model file path default:weights/best_age_estimator_resnet.h5
  • --draw, --no-draw: You can draw bounding boxes on the image. default:True

Result of age estimator on Messi image:

Result Of Identification