/Walk-or-Run-Image-Classification

This is Deep Learning project made using InveptionV3 model and Images taken from Kaggle to predict if the person in Image is Walking or Running.

Walk-or-Run Image Classification


Data

Data for this project is taken from Kaggle. It contains 500+ images for training and 100+ images for testing. This image data needs to be processed and then used to build a CNN to classifiy images. Find the data on kaggle here.

Files

The code is written in python using google colab. Find all the for the project here.

Analysis

The analysis i.e. processing of image data and model building is done using tensorflow and keras. Here I have used InceptionV3 model which is an inbuilt model in keras. InceptionV3 model gave highest 97% accuracy.

End-to-End

The webapp to test the CNN in action to classify running and walking images is built using streamlit. You can read about streamlit here. Below is the image of the webapp desing.

Screenshot