meowmeowcatcam README

The project's README from github.com/catherpiee/meowmeowcatcam. As one markdown file: README.md.

Gesture Meme Detector

Turn the camera on, make some gestures listed below to have the cat meme pop up next to you!

It will show up as:

Gestures

# Gesture How to trigger
1 Fingers Together (muehehe) Both hands up, index fingers only, tips touching
2 Devastated cat Both hands up, above the top of your head
3 Crash out Cat Both hands up beside your face, not above your head
4 Fist One hand, all four fingers curled
5 Rockstar Thumb + pinky out
6 Shhh Index finger only, tip resting on your mouth
7 One Finger Up Index finger only, held away from your face
8 Kidnap cat Any hand shape sitting where your face just was
9 i HAVE NO MONIES One open palm, all fingers extended, away from your face
10 Side Eye Just turn your head
11 Default Nothing in particular, hands down
12 Judgy cat while munching on food Look down slightly sideways
13 Huh Open your mouth and be shocked
14 O I I.A I SPIN SPIN SPIN

Meme images live in memes/. A couple of gestures pick randomly between multiple images.

Running it — desktop (Python)

Requires Python 3 and a webcam.

Easiest way: just double click Launch Gesture Meme.command. First run takes a minute to set itself up (installs everything automatically), then launches straight away. Every run after that is instant.

First time opening it: macOS will warn "cannot be opened because it is from an unidentified developer" — this is normal for any downloaded script, not specific to this one. Right-click the file → Open → click Open in the dialog that appears. You only need to do this once.

Or manually, if you prefer Terminal:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python3 gesture_meme.py

Press q or Esc in the Camera window to quit.

Running it — browser

No install needed, but the webcam API requires serving over HTTP (opening index.html directly as a file:// URL will not get camera permission). From this folder:

python3 -m http.server 8000

Then open http://localhost:8000 and allow camera access. Models load from Google's hosted MediaPipe CDN at runtime, so nothing local is needed for the browser version.

Live debug HUD

The Camera window always shows a small readout in the top-left corner:

gesture: sideEyeCat
yaw: +18.4 deg  (side-eye thr +/-15.0)

Useful for tuning the detection thresholds at the top of gesture_meme.py / app.js if a gesture is triggering too easily or not easily enough for your setup/lighting.

Project layout

gesture_meme.py   desktop version (OpenCV + MediaPipe Python tasks API)
app.js            browser version (MediaPipe tasks-vision WASM)
index.html        browser UI shell
memes/            meme images (+ one video, unused for now)
models/           MediaPipe .task model files used by the desktop version
requirements.txt  Python dependencies

meowmeowcatcam (catherpiee/meowmeowcatcam)

Searching Database…

256 ·GitHub
Meowmeowcatcam
这是什么？一个摄像头手势梗图检测器：它通过摄像头观察你的手势和头部动作，并实时在你的实时画面旁边弹出匹配的猫咪梗图。
它是桌面应用，还是在浏览器里运行的东西？两者都有——有通过可双击的 .command 文件启动的 Python 桌面版，还有直接使用摄像头 API 的纯浏览器版本（index.html + app.js），无需安装。
怎么让它跑起来？桌面版只需双击“Launch Gesture Meme.command”——首次运行会自动装好一切，之后每次都能瞬间启动。浏览器版则要把这个文件夹通过 HTTP 提供服务并打开 index.html，然后允许摄像头访问。
需要装 Python 或其他东西吗？桌面版需要 Python 3 和一个摄像头；浏览器版本地什么都不需要，因为手部跟踪模型是从 Google 托管的 MediaPipe CDN 实时加载的。
首次打开有什么要注意的坑？在 macOS 上，首次启动会警告“无法打开，因为它来自未经验证的开发者”——右键点击该文件，选择“打开”，然后确认即可，只需操作一次。浏览器版如果直接以 file:// URL 打开 index.html 也拿不到摄像头权限，必须通过 HTTP 提供服务。
它只能在 macOS 上用吗？桌面启动器是一个 macOS .command 脚本，带有 macOS 特有的“未验证开发者”警告，所以它是为 macOS 打造的；浏览器版应该能在任何有摄像头且浏览器支持摄像头的系统上工作。
它能识别哪些手势？共 14 种手势，从“Fist”、“Rockstar”（拇指+小指）和“Shhh”等手型，到“Side Eye”、“Huh”和旋转的“O I I.A I”等头/面部动作——每种手势都会触发不同的猫咪梗图。
实时调试 HUD 是干什么用的？Camera 窗口左上角始终显示一小块读数，如果某个手势在你的光线/环境下触发得太容易或不够容易，你可以据此调整 gesturememe.py 或 app.js 里的检测阈值。
能换成我自己的梗图吗？可以，梗图图片存放在 memes/ 文件夹里，而且有些手势会在该手势的多张图片中随机挑选。
怎么退出桌面应用？在 Camera 窗口获得焦点时按 q 或 Esc。
What are similar projects?itsgiving, freemocap, barehands, window-sweaters, Deep-Live-Cam, camera-to-blender, revpdf-release, facefusion
Which topics is it in?meme, gesture-control, webcam, camera, face-detection, fun, interactive, image, macos-app, app, computer-vision, html
Creator
catherpiee
Size
15.9 MB
Stars
256
Updated
CommitAug 29
Age
2mo
Commits · 90 Days