SirJAKfromSpace/MemeReview

Memes are a modern cultural phenomena that helps this generation express a vast array of sentiments albeit having short relevance life spans. Our project is a classification problem where meme images are to be recognized by :- ● Running supervised model training ● Grouping images according to visual similarity and sorted by the month/year that it was popular ● Which can then be linearly regressed to predict what type of a meme might gain popularity / be relevant.

MemeReview

Authors: Jawad Aziz Khan, Abir Rahman, Taufiqur Rahman

This contains code that can be used to visualize tens of thousands of images in a two-dimensional projection within which similar images are clustered together. The image analysis uses Tensorflow's Inception bindings, and the visualization layer uses a custom WebGL viewer.

When first running fresh, preprocessing is required. The images and required thumbnails, vector and map data can be processed by running:

python run.py "memes/*.jpg"

Or a similar 3rd path argument

Image resizing utilities require ImageMagick compiled with jpg support. For Windows search for ImageMagick and download/install using the .exe

The html viewer requires a WebGL-enabled browser. If you have a WebGL-enabled browser and a directory full of images to process, you can prepare the data for the viewer by installing the dependencies

For python 3.x, open the prompt from the project dir and run:

python -m http.server 5000

Once the web server starts, you should be able to see your results on localhost:5000. Also make sure that there is a folder named exactly "output", in the root folder before starting.