Human & Alien Image Classification

What it does

it is an ai model which classifies and differentiates between images of humans and aliens using a convolutional neural network (CNN)

How we built it

we used a convolutional neural network using Python in Jupyter Notebook. It worked by learning from training data (which were pictures of aliens and humans) and was then validated with the test set. The accuracy of the model was 0.87.

Challenges we ran into

1. found it difficult to mount Google Colab on day 1 so had to switch to Jupyter Notebook and access files locally
2. we could not use one of the TensorFlow functions due to an updating issue
3. a bug with obtaining the validation accuracy values meant that we could not plot a complete graph
4. couldn’t run the code on one of the laptops due to a bug while sampling the dataset
5. Not being able to use power apps ## What we learned How to deploy a CNN and its working; additional knowledge about coding in Python, and we improved our time management skills ## What's next for Human & Alien Image Classification

Built With

• jupyte