Breaking Frames

Our Project name is Framify, as video frames are the most fundamental unit of our project. Framify is a tool that will provide you the timestamp of the exact content you're looking for with just a prompt word. It serves by helping pinpoint these instances to you, make the task of extracting your favorite clip from a video hassle-free and thus conveniently aiding users in the creation of diverse content. Since more than 75% of people waste time watching other video parts they weren't looking for. They can use our tool to extract some specific clips from a video using a few keywords to describe the content a user is looking for. The user can also give the youtube link then Framify will fetch the video from youtube and process it. Since most videos we expect need to be processed by Framify are available on youtube, this feature will make it more convenient for the users. When the user enters a URL and a query, our AI model processes the request and returns the appropriate timestamp and other details that satisfy the given query along with the duration of the clip. Our model has an in-built OCR that can also detect the video texts that are very useful for official or formal applications. Users can get the timestamps of the part where a specific topic has been discussed and save the user's time from randomly searching for the content. It might be a student preparing a presentation, an influencer making a reel, or a professional producing commissioned content. Framify will significantly impact the education sector, cooperate sector and even be very valuable for the content creator or video editors. The tools we have used for the development are: 1. TensorFlow: Trained the deep learning model on TensorFlow using the coco2014 dataset to get the objects' labels in each frame. 2. PyTorch: the OCR module uses the Pytorch to make better detection/recognition as fast as possible. 3. Flask: used Flask to set up the server for the deep learning model 4. React: front-end is built on React.js 5. MongoDB: to store the user credentials, queries, processed video and other data 6. Express: used express.js to create an API to connect the front-end and the database Our future plans are to add a few more facial recognition models, emotion detection, summarizer using bert and automatic speech recognition models to make it more robust. We have plans to deliver it as a video streaming platform where the users can upload or stream the videos. There we will provide much more information about a video with a content extractor tool that will pinpoint the timestamp as per the query provided. Framify will significantly impact the education sector, cooperate sector and even be very valuable for the content creator or video editors as they will be getting way more content on our platform than any other.

We are a group of students with a straightforward vision. We believe that there is an amazing amount to be learned by working in a "real world" web production environment, alongside other like-minded teens. We've had an amazing journey coming to where we are today and now feel confident that we can provide you a professional and effective solution in a timely manner.