Find My Film

Inspiration

Our project is inspired from Shazam app that we use to search for songs

What it does

It takes an input of a random frame from any movie and it guesses which movie that is

How we built it

We built it using a basic CNN model. It does the job but it's a little slow. The model extracts relevant features from the dataset and performs a similarity search.

Challenges we ran into

Challenge 1: Downloading the data, movies are large files and mostly copyrighted Challenge 2: GPU Limitations Challenge 3: We did not have any front-end developer in the team, so our front-end is very basic

Accomplishments that we're proud of

1. Implementing a CNN from scratch
2. Solving an unsupervised learning problem
3. Finishing our first hackathon ## What we learned
4. Some math behind Deep Learning Models
5. Using MongoDB with Python
6. Using Streamlit to make the front-end

What's next for Find My Film

1. Improving the model, using better models
2. Scaling it over large datasets
3. Improving the speed

Built With

• mongodb
• python
• scikit-learn
• streamlit
• torch
• torchvision