Student Area

junebug-Pix2Pix

I found pix2pix so much fun to play around with. At first, I wanted to use my iPad to draw and see if anything would change with more accurate input on the user side, but I couldn’t draw inside the box with my Apple Pencil.

I tried drawing nyancat to see what pix2pix would make

I tried drawing something completely unrelated to a cat to see what would come out.

This one was a more realistic creation to see how realistic it would be.

This one was just completely random to see what would happen.

Croc with edge2shoes

I wanted to see if edges2handbags would recognize a fanny pack.

junebug-LookingOutwards04

Xander Steenbrugge’s When the painting comes to life…

Gif of a few seconds from the video

This project was an experiment in visualizing sound using AI. In the Vimeo channel’s about section, he describes his workflow and how he makes his work:

1. I first collect a dataset of images that define the visual style/theme that the AI algorithm has to learn. 
2. I then train the AI model to mimic and replicate this visual style (this is done using large amounts of computational power in the cloud and can take several days.) After training, the model is able to produce visual outputs that are similar to the training data, but entirely new and unique.
3. Next, I choose the audio and process it through a custom feature extraction pipeline written in Python. 
4. Finally, I let the AI create visual output with the audio features as input. I then start the final feedback loop where I manually edit, curate and rearrange these visual elements into the final work.

What I just appreciate about this work is the way the paintings are able to smoothly flow from one piece to another – they gradually fade into another by an element at a time rather than the whole painting. This video was so mesmerizing to watch, and the audio that was created sounded like lo-fi music and I just appreciated how in tune the visual was to the audio.

miniverse-facereadings

  1. Facial recognition algorithms used in policing are entirely based on training data. Preexisting racial bias in police provided data is contributing to weaponizing facial recognition to target poor people and minorities. Furthermore, because computer algorithms are “logical” and “correct” it canonizes biases.
  2. I like the toast fooling the facial recognition algorithms. It reminds me of “adversarial machine learning” where structured noise can completely fool a facial recognition model. It points out the flaws in these systems and makes it clear they are far from perfect.

07-faceReeadings

One idea that I thought was very interesting from the john Oliver video, although it was touched upon in a few other texts too is the idea that AI is commonly seen as purely objective. often people will view the product of some AI as proof to a statement rather than a product of inherent biased .

Another interesting point that was made in quite a few texts was the idea of universality in machine learning and AI. A lot of machines are not trained on wide cultural or racial background. This then favors people of a certain race or culture that would otherwise not be, and makes their life much easier when it comes to this piece of technology.

gregariosa-mask

 

Based on the readings last week, I wanted to create a mask that would allow people to protect their faces from facial recognition companies as a form of protest.

(I applied this animation I made on p5.JS onto the face mask, and applied the ‘Luminosity’ overlay to remove the colors.)

 

sticks-mask

 

For my mask project, I created a strange, semi-translucent mask that turns you into some sort of alien insect. 

When I was creating this mask, I was really interested in the idea of transparency, where I wanted to create a mask that masks your face, but doesn't remove all the underlying structure. The mask I created was taken from a Jello feature in Lens Studio, but distorted and transformed so that it would look more like a gas mask/face mask. I then applied a filter that gives the screen a smoky and grainy texture that works well with the translucent mask. I also added an eyeball, with the pupil being a shape of a face.

Images

Open Mouth

 

07-Mask

Crying Eyes

Do ever want people to know your crying while also covering up that ugly crying face of yours? Well now there’s a solution! Just use this snapchat lenses to let people know you’re crying while protecting them from horrifying sight of your crying face.