JerseyNet - Jersey Number Detection

JerseyNet - Jersey Number Detection

Finds jersey numbers in an image and draws a box around each one, all inside your browser so the picture never leaves your device.

Made by MHkhan50 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: JerseyNet is a small detector that finds jersey numbers in a picture and draws a box around each one it spots. JerseyNet runs inside your browser using ONNX Runtime Web. Nothing about your picture is sent to a server. It locates the numbers but does not read the digits.

Who it suits: People who study sports image analysis, or who want to see a compact detection model run entirely in a browser.

Industry | Who uses it | How they use it | Example
Sports analytics | Researchers and analysts | Locate jersey numbers in game photos | Finding the numbers in a hockey photo
Computer vision | Students and developers | Study a small detector that runs client side | Reading how the page loads the model file
Education | Teachers of machine learning | Show detection without any server | Demonstrating a model that never uploads the image

If you need to read the digits or work on other sports, you can skip it, because the author calls it a research demo.

Where to use it:

• A web browser, on the Space's page at Hugging Face
• A browser on Windows, macOS or Linux, by opening a downloaded copy of index.html

Use case | Who | What they get back
Find numbers in a photo | Anyone with a hockey image | A box drawn around each number found
Try in-browser detection | Developers | A model that runs with no upload
Study the page code | Students | A small static page with a model file and two scripts

How to set it up: Online, nothing needs installing. A local copy needs a browser and a little Python for the download.

• A web browser
• Python and the pip tool, only for the download step

1. Install the Hugging Face hub package with pip.
2. Download the Space's files into a folder.
3. Move into that folder.
4. Open index.html in a web browser.

pip install -U huggingface_hub
hf download MHkhan50/jersey-number-detection --repo-type space --local-dir jersey-number-detection
cd jersey-number-detection

How to use it: Open the Space's page, or index.html from a local copy, and upload an image. The page marks each jersey number it detects with a box. The source lists no other controls.

hf download MHkhan50/jersey-number-detection --repo-type space --local-dir jersey-number-detection

Example: You upload a photo from an ice hockey game where a player's number is clearly visible. The page draws a box around that number. It does not tell you which digits they are, and a crowded or large photo may give weaker results.

For newcomers and experts: Detection means finding where something is in a picture. To start, upload one clear hockey photo and look at the boxes. Nothing leaves your browser.

JerseyNet is a 4.7M-parameter detector in the CenterNet style. It started from random weights, with no ImageNet or COCO pretraining, and was trained on the public jersey-2 set. The training used mosaic augmentation, a GIoU box loss and 384 px inputs. On the test split it reaches AP50:95 of 0.314 and AP50 of 0.593, and 0.308 on average over three seeds. The page letterboxes the image and decodes the heatmap as in training, and the exported pipeline matches the Python results.

Upload an image in the browser
Image is letterboxed to the model input
ONNX Runtime Web runs JerseyNet locally
Boxes are drawn around the numbers

Good to know:

• Limits: its training images were small ice-hockey pictures, so results are best on clear hockey photos and poorer on other sports, big photos or busy scenes
• License: MIT
• Cost to run: none named
• Data it reaches: your image stays in the browser, and the page loads the jerseynet.onnx model file from its repository
• Not for: reading digits, other sports or production use, since it is a research demo

On Hugging Face, 13 November 2026

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

Links

• Space: https://huggingface.co/spaces/MHkhan50/jersey-number-detection
• Files: https://huggingface.co/spaces/MHkhan50/jersey-number-detection/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.