Metadata-Version: 2.1
Name: isegeval
Version: 0.1.0
Summary: A Python Library to Evaluate Interactive Segmentation Models
Home-page: https://github.com/tech-sketch/isegeval
License: MIT
Author: yasufumi
Author-email: yasufumi.taniguchi@gmail.com
Requires-Python: >=3.8.1,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Requires-Dist: numpy (>=1.23.5,<2.0.0)
Requires-Dist: scipy (>=1.9.3,<2.0.0)
Requires-Dist: tqdm (>=4.64.1,<5.0.0)
Project-URL: Repository, https://github.com/tech-sketch/isegeval
Description-Content-Type: text/markdown

# isegeval

This is a library to evaluate click-based interactive segmentation models. isegeval could evaluate
the number of click (NoC) performance of the given model on the given dataset.


## Usage

You could evaluate your model as follows. See [notebooks](./notebooks) for more detail.

```py
from isegeval import evaluate
from isegeval.core import ModelFactory


# Each item is the tuple of an image and its correspoinding ground truth mask.
dataset: Sequence[tuple[np.ndarray, np.ndarray]] = YourDataset()

# A factory of your model that you want to evaluate. The factory should implement get_model method.
model_factory: ModelFactory = YourModelFactory()

evaluate(dataset, model_factory)
```


## Installation

```bash
pip install isegeval
```

