ML Projects:

Inspiration -

Inspired by real-world applications in computer vision, we aimed to develop models that automate tasks like digit recognition and image classification.

What it does -

The first project recognizes handwritten digits using CNNs, while the second classifies images into 10 categories from the CIFAR10 dataset.

How we built it -

Both projects were built using Python, TensorFlow, and CNN architectures. We trained the models on MNIST and CIFAR10 datasets, using techniques like data augmentation and dropout to enhance performance.

Challenges -

Challenges included preventing overfitting, optimizing accuracy, and handling the complexity of small image data.

Accomplishments -

We achieved high accuracy in both models, successfully classifying handwritten digits and images.

What we learned -

We deepened our knowledge of CNNs, data preprocessing, and regularization techniques.

What's next -

Next steps include expanding the models to more complex datasets and real-world applications like medical imaging and handwriting recognition.

Built With

• cnn
• keras
• matplotlib
• mnist
• python
• tensorflow