The Gap Between Advertising and Reality

This work converts real-life room photos into commercial images. When audiences put a realistic room image into my web program, it returns a transformed image. This transformation reflects the distortions common within the advertisement industry. This piece shows the audience the distortion and transformation in commercials using the artificial network learned from the initial 800 images. The project began with the question, ‘Can Artificial Intelligence (AI) distinguish between advertising and reality?’ We used the latest machine learning technology, called Cycle GAN (a generative adversarial network), which already has classification functions.

An AI has been trained to convert images of rooms without human guidance. In the first activity, you can see how your room will look if it is advertised. Through the second activity, you can guess what the room in the advertisement actually looks like.

From Reality to Advertising

user input

Original Image

generated image

Fake Image by AI

1. You will need an unphotoshopped room image. You can take a picture of your room and use it.
2. Select and upload the image. It may take some time to reload this page with your images.



From Advertising to Reality

user input

Original Image

generated image

Fake Image by AI

1. You will need a commercial room image. You can search for keywords such as 'small room' on Google to get an image.
2. Select and upload the image. It may take some time to reload this page with your images.

Interactive Web Application

Collaboration
Kyungjin Jeong
Chanhee Cho
Year
RCA 2020
Work
Articial Intelligence
CycleGAN