Contents
Luo_Suppe2026-08-15
All items(functions/classes/consts) in this module are as follows:
| No. | item | type | introduce |
|---|---|---|---|
| 1 | math.correction.gamma | function | Gamma transform.Reference equation: Out = constant * ((data - vvmin) / vptp) ** gamma_v) * vptp. |
| 2 | math.correction.log | function | Logarithmic transformation.Reference equation: Out = log2(1 + (data - vmin) / vptp) * vptp * constant. |
| 3 | math.correction.sigmoid | function | S-shaped curve function transformation.Reference equation: Out = 1 / (1 + exp(constant * (cutoff - (data - vmin) / vptp))) * vptp. |
| 4 | math.evaluation.MAE | function | Maximum Absolute Error. |
| 5 | math.evaluation.NRMSE | function | Normalized Root Mean Square Error. |
| 6 | math.evaluation.R2 | function | Coefficient of determination. |
| 7 | math.evaluation.RMSE | function | Root Mean Square Error. |
| 8 | math.evaluation.corrcoef | function | Correlation coefficient and significance. |
| 9 | math.kernel.aggregate | function | Aggregate array. |
| 10 | math.prepro.fill_nodata | function | Interpolate missing or abnormal values. |
| 11 | math.prepro.to_numeric_array | function | Forces the input data to be converted to a numeric array, values that cannotbe converted will be modified to nan. |
| 12 | math.smooth.moving_average | function | Data smoothing based on moving average method. |
| 13 | math.smooth.savitzky_golay | function | Data smoothing based on Savitzky-Golay method. |
| 14 | math.stretch.hist_equalize | function | Apply non-linear contrast stretching, where the values are distributed within the bit depth range. |
| 15 | math.stretch.max_min | function | Use linear stretching with the minimum and maximum output pixel values as the endpoints of the histogram. |
| 16 | math.stretch.percentage | function | Truncates the array based on the percentage interval entered.Then stretch it to the specified range. |
| 17 | math.stretch.std | function | Standard deviation stretch. This method is used to emphasize how much feature values vary from the mean value;it is best when used on normally distributed data. |