Machine Learning, Computer Vision,

STEM Thesis, Individual Project

Fall 2019 - Present (ongoing research)

This a year-long thesis project investigates a new preferable interaction between humans and machines through the AI 'eyes', exploring new possibilities for people to live with machine vision and other intelligence. The project aims to assist individuals in branding themselves to various algorithms, proposing new machine-coauthored aesthetics and blending up it with people's mundane life, which leads to a brand new lifestyle.

Screenshot 2020-02-18 at 10.30.58.png
Screenshot 2020-02-18 at 10.38.15.png

Image generated by the ACE algorithm interprets high-confidence concept segments from the original image. This collage collected the patches selected by the algorithm in a cute cat database.

Screenshot 2020-02-18 at 10.31.21.png
Screenshot 2020-02-18 at 10.31.36.png

How might we find a way to put the discrete world together since we are living with the segmented machine vision and keep getting fragmented information processed by the machine? 

Validating the mug I designed to see whether the algorithm can recognize it or not.

Screenshot 2020-02-18 at 10.28.37.png

The impressive result shows that the mug I designed can be detected as eight different objects rather than just a mug. Compares to a regular looking cup, the algorithm has less confidence to recognise and mask my mug.

paper-bag-02 copy_s.png

A package interpreted by machine as vegan food.

Exhibition during the WIP show in the Wind Tunnel at Media Design Practices.