Module: sarpyx.utils.metrics

File: sarpyx/utils/metrics.py

Performance Metrics for Complex-Valued SAR Image Super-Resolution. This module provides comprehensive evaluation metrics for Synthetic Aperture Radar (SAR) super-resolution algorithms. SAR super-resolution aims to enhance spatial resolution while preserving both amplitude (magnitude) and phase information. The metrics are organized into three main categories: 1. **Magnitude Fidelity Metrics** (Amplitude Domain): - Mean Squared Error (MSE) and Root MSE (RMSE) - Peak Signal-to-Noise Ratio (PSNR) - Structural Similarity Index (SSIM) - Amplitude Correlation Coefficient (CC) 2. **Phase Accuracy Metrics** (Phase Domain): - Mean Phase Error (MAE/RMSE of Phase) - Complex Correlation Coefficient (Interferometric Coherence) - Phase-Only Correlation (Phase Consistency) 3. **SAR-Specific Structural Metrics**: - Equivalent Number of Looks (ENL) - Resolution Gain (Spatial Resolution Improvement) All metrics accept complex-valued numpy arrays representing SAR images and provide quantitative assessment of super-resolution quality for both research and operational applications. References: IEEE Geoscience and Remote Sensing literature for SAR image quality assessment, Remote Sensing journal publications on SAR super-resolution, and IEEE TGRS articles on interferometric coherence and SAR image processing.

Exported Symbols (__all__)

No explicit __all__ list. Public symbols inferred from implementation.

Public Functions (11)

mse_complex(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import mse_complex

result = mse_complex(ref=<ref>, pred=<pred>)

rmse_complex(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import rmse_complex

result = rmse_complex(ref=<ref>, pred=<pred>)

psnr_amplitude(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import psnr_amplitude

result = psnr_amplitude(ref=<ref>, pred=<pred>)

ssim_amplitude(ref: np.ndarray, pred: np.ndarray, window_size: int=7, return_components: bool=False) -> Union[float, Tuple[float, float, float, float]]

from sarpyx.utils.metrics import ssim_amplitude

result = ssim_amplitude(ref=<ref>, pred=<pred>)

amplitude_correlation(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import amplitude_correlation

result = amplitude_correlation(ref=<ref>, pred=<pred>)

phase_error_stats(ref: np.ndarray, pred: np.ndarray) -> Tuple[float, float]

from sarpyx.utils.metrics import phase_error_stats

result = phase_error_stats(ref=<ref>, pred=<pred>)

complex_coherence(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import complex_coherence

result = complex_coherence(ref=<ref>, pred=<pred>)

phase_coherence(ref: np.ndarray, pred: np.ndarray) -> float

from sarpyx.utils.metrics import phase_coherence

result = phase_coherence(ref=<ref>, pred=<pred>)

enl(intensity_image: np.ndarray) -> float

from sarpyx.utils.metrics import enl

result = enl(intensity_image=<intensity_image>)

resolution_gain(orig: np.ndarray, sr: np.ndarray, threshold: float=0.5) -> float

from sarpyx.utils.metrics import resolution_gain

result = resolution_gain(orig=<orig>, sr=<sr>)

evaluate_sar_metrics(ref: np.ndarray, pred: np.ndarray, window_size: int=7, threshold: float=0.5) -> dict

from sarpyx.utils.metrics import evaluate_sar_metrics

result = evaluate_sar_metrics(ref=<ref>, pred=<pred>)

Public Classes (0)

No public classes detected.