Speaker

Emad Ehsan

Emad Ehsan

Google Developer Expert

Emad is the youngest Google Developer Expert from Pakistan. He is Chief Technology Officer at Traverous - a travel journal and video making app for travelers.

He has been a Top Trending GitHub Developer and contributes regularly to open source. He has spoken at Tech Conferences in UK, Denmark, US, Pakistan, & Malaysia.

Area of Expertise

  • Information & Communications Technology
  • Transports & Logistics
  • Agriculture, Food & Forestry
  • Environment and Cleantech
  • Travel & Tourism

Topics

  • Google Cloud
  • Machine Learning
  • Operations Research
  • Algorithms

Building custom Deep Learning model to recognize British Foods with Cloud AutoML & no code

Google Cloud offers 2 tremendous tools to get you up to speed with building AI powered Apps. Vision API and AutoML. Vision API is great at recognizing the contents of an image but it's not very flexible. Let's say you want to build app for Local Restaurants, or a Food Image Search app for tourists that tells the names of food in images as known by locals.

For example Bacon Sandwitch is known as bacon butty in UK, or bacon bap in NZ. Scottish Haggis looks very similar to Qeema (crushed beef dish) of Pakistan and you'd definitely want your app to recognize and distinguish between the two.

To build these type of custom and localized experiences (apps) for your users, AutoML is the best tool for you. With AutoML, you don't even need to write any code to make this custom DL model.

In this project, we will use a few hundred images from the internet to train our custom Deep Learning model to recognize British Foods by their original names, with no code. We'll also discuss how to integrate this model in a mobile or web app.

GDG DevFest UK & Ireland 2020

October 2020 London, United Kingdom