Skip to content
Source code used for the talk: Solving “Where’s Wally?” with Turi Create
Swift Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Classes
Resources
Turi Create Resources
WallyML.xcodeproj
.gitignore
LICENSE
README.md

README.md

Solving “Where’s Wally?” with Turi Create (iOS)

Source code used for my talk: Solving “Where’s Wally?” with Turi Create

Overview

This project uses the Machine Learning framework Turi Create in order to solve the game “Where’s Wally?”.

The support of this talk can be found here.

The video of the talk can be found here.

As explained during this talk, solving "Where's Wally?" using Machine Learning is a tough challenge for multiple reasons:

  • Not a lot of data available to train the model
  • Not a lot of good resolution data available
  • Wally size on each picture is really small
  • The complicity of the whole picture. Many colors and shapes used to make the real Wally difficult to find

As a Machine Learning novice, the goal of this project was not to have a bullet proof Wally finder, but more to see the possibilities of the Machine Learning technology and the Turi Create framwork with an interesting and fun project.

Create the Turi Create model on your own machine

If you want to create the Turi Create model used for this project in your own machine, the original images are also available in the repository.

Check this other related Github repository in order to generate the training/testing images and the annotations.csv file.

When it is done, you can know start to create the model.

Here are the simple steps to create it:

  1. Check the System Requirements to check if your machine and/or Python version can run Turi Create
  2. Follow the guide to install Turi Create (I also recommend using virtualenv to use, install, or build Turi Create.)
  3. In the Turi Create Resources folder, run python WallyML.py

TensorFlow users

If you are a TensorFlow user for Object Detection and want to try Turi Create, I also open-sourced TFannotations2TC, a little Python script to convert the TensorFlow annotations csv file into a Turi Create annotations csv file. Nothing fancy here, but can help some people.

References

Turi Create documentation: https://github.com/apple/turicreate

Turi Create (Object Detection): https://apple.github.io/turicreate/docs/userguide/object_detection/

Repository containing "Where's Wally?" images for training: https://github.com/tadejmagajna/HereIsWally

Further reading

Finding Waldo Using Semantic Segmentation & Tiramisu: https://hackernoon.com/wheres-waldo-terminator-edition-8b3bd0805741

Tensorflow project that includes a model for solving Where's Wally puzzles: https://github.com/tadejmagajna/HereIsWally

How to Find Wally with a Neural Network: https://towardsdatascience.com/how-to-find-wally-neural-network-eddbb20b0b90

Contact

If you have a question or any comment, feel free to open an issue or to DM me on @florianldt.

You can’t perform that action at this time.