Inspiration

The basic inspiration is to perform deep analysis on the social media pictures that we post to get insights which have huge application in the field of Healthcare (doctors can know if the patient is healthy, determine mental conditions using Color palette determination), business application to determine the locations a person is visiting and send him/her offers from local merchants or even for the platform (Instagrams's) owners to detect any indecent pictures in the real time.

What it does

This software leverages Instagram API for login and allows users to download images in real time which are then analyzed using GCP Vision API to determine key features like objects, description etc in the picture.

How I built it

  • I used Instagram API for signing in and downloading top 5 images.
  • Then I used XAMPP server for coding the API and developing the landing page. *I created a Google Cloud Platform Virtual Machine environment using Keygen Putty and also used PHP to code the Vision API
  • Finally, I used Python in the Virtual Machine I created to determine the outcome of Google Cloud Platform's API

Challenges I ran into

After running the Google Cloud Platform API using PHP on the localhost, the code gave errors, hence I used Virtual Machine environment to verify results and have a plan B.

Accomplishments that I'm proud of

setting up the virtual environment. I have deep interest in blockchain applications, Deep Learning and app development. Almost all of them require SSH key authorization. Using Puttygen, all of this became really easy. I struggled a bit as I was working alone, hence getting help was difficult but the mentors have been a great help.

What I learned

Learnt about Putty, puttygen, SSH key deployment, setting up a virtual machine, Nano and Vim editors along with Google Cloud Platform and numerous applications it has. Before doing this project I thought IBM Watson is the best tool for AI but I think Google Cloud Platform is super powerful as it eliminates the requirement to train the system.

What's next for Deep Vision 3001

  • Firstly, this is going to be my final year project in the Big Data and AI class.
  • Develop an image recommender system which uses GCP Vision API to detect objects and return similar results by returning searches based on the objects detected.
  • Contribute examples and demos to the GCP platform by posting them on my Github account
Share this project:

Updates