PySODMetrics Documentation

Welcome to PySODMetrics - A simple and efficient implementation of SOD metrics.

PyPI version Downloads

Overview

PySODMetrics is a Python library that provides simple and efficient implementations of metrics for evaluating salient object detection (SOD), camouflaged object detection (COD), and medical image segmentation tasks.

Key Features:

  • Based on numpy and scipy for fast computation

  • Verified against the original MATLAB implementations

  • Simple and extensible code structure

  • Lightweight and easy to use

Note

Our exploration in this field continues with PyIRSTDMetrics, a project born from the same core motivation. Think of them as twin initiatives: this project maps the landscape of current evaluation, while its sibling takes the next step to expand upon and rethink it.

Contents

User Guide

API Reference

Supported Metrics

PySODMetrics supports a comprehensive set of evaluation metrics:

  • MAE - Mean Absolute Error

  • S-measure (\(S_m\)) - Structure Measure

  • E-measure (\(E_m\)) - Enhanced-alignment Measure

  • F-measure (\(F_\beta\)) - Precision-Recall F-measure

  • Weighted F-measure (\(F^\omega_\beta\))

  • Context-Measure (\(C_\beta\), \(C^\omega_\beta\))

  • Multi-Scale IoU - Multi-scale Intersection over Union

  • Human Correction Effort Measure

  • And many more classification metrics (BER, Dice, Kappa, Precision, Recall, etc.)

See Supported Metrics for detailed descriptions of all supported metrics.

Indices and tables