okay so omoggle/mogwars was cool but this is cooler:

i noticed that those “backends” only ran on a static frame of your face and only measured landmarks. I decided to make one which bases it off of a 3D scan as well as texture information to estimate skincare.

After some research, I found 3DDFA, a small model which can reconstruct a 3D model of a face from photos.

the app first takes 7 poses by asking you to move your camera and then passes them all into the model. (120x120x3b tensor). The model gives 62 parameters per pose, and then the app takes the median identity-shape parameters across all views -> BFM2009 decoder -> 68 points used to calculate all the structural metrics.

then it uses the texture of the face that’s not in the eyes or mouth and samples them for clarity, eveness, texture, exposure, sharpness, etc. takes the median of that and then calculates a skin value. exact weights for them are found in the code.

Mogged

sorry for the lackluster devlog quality, my >4k char devlog was js deleted and i’m pissed

inspo

a very chopped aquantience couldn’t believe how chopped he was and I wanted to prove it to him. however, he didn’t want to go into an omoggle match w me so I needed something that supported image upload

motivation

popular websites that do an omoggle-style ranking either are not accurate to omoggle or send your photo off-device to a vLLM or server, which I don’t want

criteria

• both image and photo support
• privacy first all local on-device scoring
• accurate to the two popular mogging platforms I found, mogwars and omoggle

research

• found an excellent (althrough made by ai) writeup on omoggle. This shows that Omoggle uses Google’s Mediapipe FaceLandmarker. I knew the data fed into an algorithim but I didn’t know how the facial data was turned into a score
• so I got a codex agent to analyze the bundled JS of omoggle and mogwars to reverse-engineer the scoring mechanism. It came back with the following:
  • that both websites use the same metrics but process them differently into a score
  • the way metrics (symmetry, jawline, etc.) are calculated from the landmarks
  • and how each platform uses them

for sake of brevity (character limit) I will not be describing exactly how each platform works nor how the metrics are calculated in specific, for those interested, check out src/lib/scoring. I promise I’m not larping, I actually understand how it works lol.

implementation

I then took all of that as well as some Google-given examples for the landmarking API and created a new NextJS + Shadcn + Tailwind v4 project. I first wrote the landmarking system and then the scoring system based on the output of the landmarker. After, I made a temporary UI with google stitch and got codex to wire everything together.

demo

available at https://mogged.eli.best