Blog

Butterfly StyleGAN2

Objective: Use Runway to train my own StyleGAN2 model using butterfly images from Flicker. I first scraped 1919 images from Flicker using the keyword “butterfly”. Then, I uploaded the dataset to Runway. I used StyleGAN2 with the Objects pre-trained model and 3000 steps to train my butterfly model. I applied center crop to all my images to make […]

Read More Butterfly StyleGAN2

Dr. Seuss + Markov Chains

Objective: Generate new stories and poems based on a Markov chain trained with Dr. Seuss Using Daniel Shiffman’s example code on Markov chains by words, I used Dr. Seuss’ book Oh, the Places You’ll Go! as my input. Dr. Seuss is one of my favorite childhood authors so I was curious to explore what new […]

Read More Dr. Seuss + Markov Chains

Spade COCO Generated Video

Objective: Create a p5 sketch using Spade-COCO’s recognized colors to generate a photorealistic beach scene First, I created a p5 sketch to serve as a semantic image for the Spade COCO model in RunwayML. I used 6 colors associated with following labels: sand, sea, sky, tree, leaves, and person. The person diving in the sketch […]

Read More Spade COCO Generated Video

Style Transfer

Objective: Use a pre-trained model to create a style transfer from one image to another Using ml5.js library and the style transfer pre-loaded model, I created the following image above. I used the example code from Yining Shi to build my program. The style I wanted to apply on my picture is The Great Wave […]

Read More Style Transfer

Packet Analysis

To get comfortable with Wireshark and how it works, I performed 5 different captures and analyzed the traffic. Here are my most interesting findings: Notable endpoints: Amazon Google Microsoft Edgecast [content delivery network (CDN)] Cloudflarenet [web-infrastructure and website-security company] Github Automattic Inc [Wordpress] Fastly [cloud computing services provider] After applying the filter tcp.port == 80 […]

Read More Packet Analysis

Sound Classifier + Arduino

Objective: Use the sound classifier model from ml5.js to control a 7 segment LED display connected to an Arduino Uno To bring hardware and p5.js together, I first focused on building my circuit. I found this reference diagram below and used it to connect my LED display model # 350074. Next, I worked on writing […]

Read More Sound Classifier + Arduino

Face Matrix

Objective: Use the Face API machine learning model to create a Face Matrix! My project detects a user’s face using the Face API from ml5.js and draws real-time face landmarks. This was more of an exercise for me to practice working with the Face API and understand how it works. I wanted to experiment making each […]

Read More Face Matrix

Dancing Pumpkin Character

Objective: Use the Posenet machine learning model to recreate my favorite halloween gif! It is October! Which means Halloween is fast approaching! In the holiday spirit, I decided to recreate my favorite halloween gif: the dancing Halloween man. My friends and I always send this gif to each other whenever we are feeling down. My […]

Read More Dancing Pumpkin Character

2020 Presidential Debate Analysis

Objective: Visualize a word count of each participant in the presidential debate on September 30, 2020. With the United States presidential election coming up, I wanted to explore and analyze the transcript of the debate this past week. First, I downloaded a transcript from rev.com as a .txt file. Then, I created 3 dictionaries in […]

Read More 2020 Presidential Debate Analysis

Corpora + BodyPix

Objective: Explore with the BodyPix machine learning model and p5.js. I wanted to create a virtual background using the BodyPix machine learning model from ml5.js. The background consists of a blue mask with a list of the most popular programming languages from Corpora. It is a relatively short and simply program to keep me inspired and encourage […]

Read More Corpora + BodyPix