Face Detector
Detect faces and eyes, with a box around each.
What does the face detector do?
The freeimgtool face detector finds the location of every front-facing face in an uploaded image, draws a green box around each one, and looks for eyes inside each face box. It uses OpenCV's Haar cascade classifiers, the classical face-detection technique that ships with OpenCV. The output includes both an annotated image and a table listing each face's coordinates, size, and eye count.
This is not a deep face recognition model — it cannot tell who is in the photo. It only marks where faces are. For recognition tasks a heavier model is needed.
How to use it
- Click the upload box and pick a photo (up to 5 MB).
- Click Detect faces.
- Read the face count and the per-face table.
- Download the annotated image with face and eye boxes.
Why use this face detector
Useful for quick auditing: counting attendees in a group photo, sanity-checking a dataset, building a privacy compliance check, or as a teaching aid for computer vision courses. The Haar approach is honest about its limits — it handles upright frontal faces best, struggles with profiles and heavy occlusion.
Who uses it
Event organisers counting attendees. Photographers checking group shots. CS students studying computer vision. Privacy engineers prototyping blurring pipelines. Anyone curious how many faces a classical detector finds in their photo.
Frequently asked questions
Is this free?
Yes. No signup, no premium tier. Display ads support the site.
What is the maximum upload size?
Five megabytes per photo. Compress first if your file is larger.
Why did it miss a face?
Haar cascades work best on upright, frontal faces with good lighting. They struggle with profiles, faces tilted past about 20 degrees, sunglasses, and very small faces. Re-shoot frontally if you can.
Does this recognize who is in the photo?
No. The detector only locates faces. It does not store, identify, or match identities.
Is my photo stored?
Uploads are kept only long enough to return the annotated image link and are removed automatically. Privacy policy has the details.
What output format do I get?
JPG at good quality, with green boxes around every face and yellow boxes around the first two eyes inside each face.
Does it work on a phone?
Yes. The upload box opens the camera roll on most phones directly.
What about full face landmark detection?
The current build returns bounding boxes only. Full 68-point or mesh landmark detection is on the roadmap for the upgraded server.