Rabbit Pain Identifier

Rabbit Pain Identifier

Screens a single rabbit photo for signs of pain by measuring 23 facial landmarks and labelling it Relaxed or Pain with a probability. A research prototype for rabbit owners, vets and students, not a diagnosis.

Made by sheildo on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Space looks at one photo of a rabbit and estimates whether the animal seems relaxed or in pain. It finds 23 landmarks on the face, corrects for head tilt, measures facial geometry and passes the numbers to a classifier. The result is a label with a probability score.

Who it suits: Rabbit owners, veterinary students and researchers who want to see how photo-based pain screening works.

Industry | Who uses it | How they use it | Example
Veterinary research | Animal welfare researchers | Study facial geometry as a pain signal | Run a photo and read the probability score
Veterinary education | Students | Learn how a screening model is built | Open the sample photos and compare the labels
Pet care | Rabbit owners | Get a first, informal look at a photo | Upload a clear front-facing photo of their rabbit
Software | ML developers | Study a small landmark and KNN pipeline | Read app.py next to the saved model files

If you need a diagnosis, ask a vet. The page says it is a research prototype and not a diagnosis.

Where to use it: It runs in two places.

• A web browser, on the Space's page at Hugging Face with nothing to install
• A local copy on Windows, macOS or Linux, which needs Python 3.12

Use case | Who | What they get back
Screen one photo | Owners and vets | A Relaxed or Pain label with a probability
Try the samples | Students | Results for the lop ear, pain and relaxed sample photos
Study the pipeline | Developers | The model files and the app code in one repository

How to set it up: Using it online needs nothing. A local copy needs a few things.

• Python 3.12 and pip
• The packages listed in requirements.txt
• The system packages listed in packages.txt, installed with apt-get

1. Install the Hugging Face tools.
2. Download the Space into a folder.
3. Move into that folder.
4. Install Gradio and the packages from requirements.txt.
5. Start the app.

pip install -U huggingface_hub
hf download sheildo/Rabbit-Pain-Identifier-Model --repo-type space --local-dir Rabbit-Pain-Identifier-Model
cd Rabbit-Pain-Identifier-Model
pip install gradio==5.49.0
pip install -r requirements.txt
python app.py

How to use it: Open the page and add a single photo of a rabbit. The page returns a Relaxed or Pain label and the raw probability. For screening, where you want to catch every possible pain case, use a threshold of 0.4. For triage, where you want fewer false alarms, use 0.5.

Example: You upload a clear photo of a rabbit with upright ears. The page marks the facial landmarks and reports a label with a probability, which you compare against the 0.4 or 0.5 threshold that suits your goal.

For newcomers and experts: Start with one of the sample photos in the repository to see how a result looks. Remember that a photo is a rough guide and a clinician has the final say.

The pipeline uses a YOLOv8n-pose model to find 23 landmarks. It turns the landmarks around the point between the eyes until the eyes sit level, and it scales every measurement by how far apart the eyes are. For floppy-eared rabbits, an ear drop check turns off the ear features. A KNN classifier then predicts the label.

YOLOv8n-pose finds 23 facial landmarks
Landmarks are rotated to level the eyes
Features are scaled by the eye distance
A KNN classifier scores Relaxed or Pain

Good to know: The model was trained on one curated set of 620 images, and the author says outside validation is still pending.

Title: Accuracy of the classifier on its checks
Test set: 0.895 accuracy
Validation set: 0.871 accuracy
5-fold cross-validation: 0.820 accuracy
Majority baseline: 0.659 accuracy

• Each photo takes about 2 to 6 seconds on the free CPU tier
• Results come from a research prototype and are not a veterinary diagnosis
• License: MIT
• Cost to run: none named
• Data it reaches: the photos you upload
• Not for: veterinary diagnosis, which belongs with a clinician and the Rabbit Grimace Scale

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: mit

Links

• Space: https://huggingface.co/spaces/sheildo/Rabbit-Pain-Identifier-Model
• Files: https://huggingface.co/spaces/sheildo/Rabbit-Pain-Identifier-Model/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.