Create Artistic Effect by Stylizing the Image Background

[Margaret Maynard-Reid]

Hi everyone,

Have you ever wondered how the background stylization works in popular apps such as Google Meet, Zoom or Microsoft Teams? Or style filters in Snapchat or in Photoshop new features? If so, read on! 

In our latest end-to-end TensorFlow Lite tutorial, we combined two TensorFlow Lite models ( segmentation and style transfer) to create an artistic effect for the image background.

Compared to our previous end-to-end tutorials, this one is more closely aligned with the real-world scenario of needing to chain together multiple models.

It also showcases many new TFLite features including the benchmarking tool, the (segmentation) task library and ML model binding etc.

This tutorial is divided into four parts -  

• Part 1: Project Intro (this one), Margaret Maynard-Reid

• Part 2: TensorFlow Lite models, Sayak Paul

• Part 3: Android implementation, George Soloupis

• Part 4: iOS implementation (coming soon…), Patrick Haralabidis

Consider retweeting the project to help us reach more audiences.   

This project is part of the on-going effort to create End-to-End TensorFlow Tutorials with sample code. It’s a collaboration by ML GDEs and members from the community.

We would like to thank the TensorFlow Lite team (in particular Khanh & Lu) for their continuous support.

Many thanks,

Margaret / Sayak / George / Patrick

[Sayak Paul]

Finally, it's OUT! Yay! 🤪

@Jason Mayes I think it might be cool for the TFJs working group to work on a similar browser extension? If needed we are always happy to walk you through our approach. 

Sayak Paul | sayak.dev

[Lu Wang]

+tflite-team 

Congratulations to the team! Very exciting! It's a super great showcase of TFLite features!

[Soonson Kwon]

Big congrats to this virtual team!

I am particularly happy to see a collaboration among a diverse group of people.

(Technology background, geographic... etc etc.)

A continued great example of a great collaboration project indeed.

[Justin Hong]

Awesome! Congrats to everyone who made this work!

--
You received this message because you are subscribed to the Google Groups "tflite-team" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tflite-team...@google.com.
To view this discussion on the web visit https://groups.google.com/a/google.com/d/msgid/tflite-team/CA%2BZ%2BUYM6vh0Vv1hiiu5aw8kYYpXJdNJp1VOyQhqa6P9rF%2Bf%3Dtg%40mail.gmail.com.

[Khanh LeViet]

Huge thanks to George, Margaret, Sayak and Patrick for making this amazing project happen!

I have also asked TensorFlow Twitter account to retweet Margaret's tweet :)

 

--

    

Le Viet Gia Khanh (カン)

TensorFlow Developer Advocate

mail: khan...@google.com

tel: +81 3-6384-6867 

Shibuya Stream 

3-21-3 Shibuya, Shibuya-ku, Tokyo

150-0002, Japan

[Shuangfeng Li]

This is super cool to talk about a real case that is widely used in production right now! Congratulations for the great work!

On Sat, Nov 28, 2020 at 4:33 PM Shweta Bhatt <shweta...@gmail.com> wrote:

This is pretty cool! Congratulations team :)

Best,

Shweta

You received this message because you are subscribed to the Google Groups "ML GDE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ml-gde+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ml-gde/CAMDTK-6vM55ZujTNVSSJB%3DqrbtOqekVs_F1wvMRJUSmfDE0QVQ%40mail.gmail.com.