Metadata-Version: 2.5
Name: seg-studio
Version: 0.9.8.post2
Summary: Pointer package for Seg-Studio, an open-source local GUI for image segmentation — SAM-assisted annotation, PyTorch training and ONNX / CoreML export.
Project-URL: Homepage, https://segmen-pixel.github.io/seg-studio/
Project-URL: Documentation, https://segmen-pixel.github.io/seg-studio/
Project-URL: Source, https://github.com/segmen-pixel/seg-studio
Project-URL: Download, https://github.com/segmen-pixel/seg-studio/releases/latest
Project-URL: Changelog, https://github.com/segmen-pixel/seg-studio/blob/main/CHANGELOG.md
Project-URL: Issues, https://github.com/segmen-pixel/seg-studio/issues
Author: segmen-pixel
License-Expression: Apache-2.0
License-File: LICENSE
Keywords: annotation,annotation-tool,computer-vision,coreml,deep-learning,gui,image-annotation,image-segmentation,labeling,machine-learning,object-counting,offline,onnx,openvino,pytorch,sam,segment-anything,semantic-segmentation
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: X11 Applications :: Qt
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: MacOS :: MacOS X
Classifier: Operating System :: Microsoft :: Windows
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Multimedia :: Graphics :: Editors
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Scientific/Engineering :: Image Processing
Classifier: Topic :: Scientific/Engineering :: Image Recognition
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# Seg-Studio

**An open-source, all-in-one local GUI for image segmentation — annotate, train and export on your own machine.**

[Documentation](https://segmen-pixel.github.io/seg-studio/) ·
[Source](https://github.com/segmen-pixel/seg-studio) ·
[Download](https://github.com/segmen-pixel/seg-studio/releases/latest) ·
[日本語](https://segmen-pixel.github.io/seg-studio/ja/)

> **This PyPI package is a pointer, not the application.** Seg-Studio is a
> desktop app distributed as a ZIP from the GitHub releases page, so
> `pip install seg-studio` only installs a `seg-studio` command that prints
> the download link. The install instructions are below.

---

## What Seg-Studio is

Seg-Studio is a workbench for semantic segmentation that runs entirely on your
own machine. You annotate images with SAM-assisted tools, train PyTorch models
with live monitoring, and export the result to ONNX or CoreML for edge
deployment. There is no cloud account, no upload step and no telemetry — the
images and the trained weights stay on the machine you run it on.

Alongside semantic segmentation it can count individual objects, reusing the
masks you already drew rather than asking for a second round of annotation.

### Features

- **SAM-assisted annotation** — click-to-segment, brush and polygon tools, class
  management and autosave. Masks are plain indexed PNGs, so nothing is locked in.
- **Training without writing code** — the recipe (backbone, epochs,
  augmentation, learning rate) is selected for you, and every field stays
  editable if you want to override it.
- **Live monitoring and evaluation** — loss and metric curves during training,
  then per-image F1 / precision / recall / IoU, confidence distributions and a
  heatmap view.
- **Object counting** — instance counts derived from the masks you annotated.
- **Export for deployment** — ONNX, OpenVINO IR and CoreML, so a model trained
  here runs on an edge device, an iPhone or a PC without Seg-Studio installed.
- **Runs offline** — Windows and macOS, CPU or CUDA GPU, Apple Silicon via MPS.

## Installing the application

1. Download the ZIP from the
   [releases page](https://github.com/segmen-pixel/seg-studio/releases/latest).
2. Run the install script for your platform — `install-windows.bat` or
   `install-macos.sh`.
3. Run `start-windows.bat` or `start-macos.sh`.
4. Open <http://localhost:8002/ui/> in a browser.

Python 3.10 or newer is required. The
[first-run walkthrough](https://segmen-pixel.github.io/seg-studio/docs/first-run-manual.html)
takes you from there to a trained model in about ten minutes, and the
[troubleshooting guide](https://segmen-pixel.github.io/seg-studio/docs/troubleshooting.html)
covers the common install problems.

## Status and licence

Current release: **v0.9.8 (beta)**, licensed under Apache-2.0. Third-party
components and their licences are listed in
[THIRD_PARTY_NOTICES](https://github.com/segmen-pixel/seg-studio/blob/main/THIRD_PARTY_NOTICES.md).
