Extracting frames from videos sounds simple… until you actually need:
Custom frame intervals
Privacy protection (faces, license plates)
Clean datasets for AI or photogrammetry
A usable UI (not just scripts)
So I built a desktop tool to solve all of that.
👉 VID2IMG Free — a Python-based video-to-image extraction app with built-in anonymization and a modern UI.
🚀 What This Tool Does
At its core, VID2IMG uses OpenCV to process video frames and export them as images.
But I added a few features that make it more practical for real-world use:
🎥 Frame Extraction Engine
Supports MP4, AVI, MOV
Configurable frame interval
Efficient sequential processing
if idx % frame_interval.get() == 0:
cv2.imwrite(path, frame)
🔒 Anonymization (Computer Vision)
Using Haar cascades from OpenCV:
Face detection → Gaussian blur
License plates → pixelation
faces = face_cascade.detectMultiScale(gray, 1.2, 5)
frame[y:y+h, x:x+w] = cv2.GaussianBlur(roi, (51, 51), 0)
This makes it useful for:
Privacy-safe datasets
Street footage processing
📸 Photogrammetry Mode
When working with tools like Meshroom or RealityCapture:
Sequential filenames (frame_000001)
Lossless PNG output
Consistent resolution
cv2.imwrite(path, frame, [cv2.IMWRITE_PNG_COMPRESSION, 0])
🌍 360° Video Safe Mode
Handles equirectangular video without breaking geometry (important for VR / 360 workflows).
🖥️ UI with Tkinter + ttkbootstrap
Instead of CLI-only, I built a full desktop UI:
Modern theme (ttkbootstrap)
Progress bar
Live logging (thread-safe)
def ui(func, *args, **kwargs):
app.after(0, lambda: func(*args, **kwargs))
Threading keeps the UI responsive while processing video.
⚙️ Architecture Overview
Frontend: Tkinter + ttkbootstrap
Processing: OpenCV (cv2)
Concurrency: Python threading
Packaging: PyInstaller (EXE build)
📦 Download (Windows EXE)
I packaged it into a standalone executable:
👉 https://github.com/rogers-cyber/VID2IMG-Free/releases/tag/v1.0.0
No Python required — just run the .exe.
⚠️ Notes on Distribution
Since this is packaged with PyInstaller:
Windows SmartScreen may warn (unsigned EXE)
Some antivirus tools may flag it (false positive)
This is expected for indie-built tools.
📢 Why This Version Has Ads
To keep the tool free, I added:
Small rotating banner ads
No tracking
No background processes
It’s a simple way to support development without paywalls.
💎 Pro Version
I also built a Pro version (no ads) with a cleaner experience and future upgrades.
👉 https://matetools.gumroad.com
🧠 Use Cases
This tool is useful for:
AI / ML dataset generation
Privacy-compliant video processing
Drone / action camera workflows
3D reconstruction pipelines
🛠️ Tech Stack
Python
OpenCV
Tkinter
ttkbootstrap
Pillow
🔥 Lessons Learned
A few interesting takeaways while building this:
Thread-safe UI updates in Tkinter are critical
OpenCV Haar cascades are still useful for lightweight detection
Packaging Python apps for Windows is harder than expected
Small UX details (like logs + progress bars) matter a lot
⭐ Final Thoughts
This started as a personal utility, but turned into a full tool others can use.
If you're working with video processing or datasets, give it a try.
Feedback is welcome — I’m planning more features soon 🚀
Thanks for reading 🙌
Top comments (0)