/theres-wally

Primary LanguageJupyter Notebook

There's Wally

An experiment in using a Convolutional Neural Network to find Wally in Where's Wally images.

The code is based on @HTuennermann's code for detecting and locating cars, available here: https://github.com/HTuennermann/Vehicle-Detection-and-Tracking

Installing dependencies

$ pip install -r requirements.txt

Starting the Jupyter notebook server

$ jupyter notebook

Training the network

Run the train notebook.

You'll need a bunch of images that I can't share on GitHub for copyright reasons.

Using the pre-trained network

  1. Save your Where's Wally image as test.jpg in the root folder of this repo.

  2. Run the find-wally notebook.