Skip to content

ColorCorrectionAnalyzer class

ColorCorrectionAnalyzer

Analyzer for benchmarking color correction methods.

This class combines multiple correction and detection techniques to analyze the performance of color correction models by comparing patches and overall image differences.

Parameters:

Name Type Description Default
list_correction_methods list of tuple[LiteralModelCorrection, dict]

A list of tuples, where each tuple contains a correction method identifier and its parameters.

 required
list_detection_methods list of tuple[LiteralModelDetection, dict]

A list of tuples, where each tuple contains a detection method identifier and its parameters.

 required
use_gpu bool

Flag to indicate whether GPU is used, by default False.

 False
Source code in color_correction/services/correction_analyzer.py 
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
class ColorCorrectionAnalyzer:
    """
    Analyzer for benchmarking color correction methods.

    This class combines multiple correction and detection techniques to analyze
    the performance of color correction models by comparing patches and overall image
    differences.

    Parameters
    ----------
    list_correction_methods : list of tuple[LiteralModelCorrection, dict]
        A list of tuples, where each tuple contains a correction method identifier
        and its parameters.
    list_detection_methods : list of tuple[LiteralModelDetection, dict]
        A list of tuples, where each tuple contains a detection method identifier
        and its parameters.
    use_gpu : bool, optional
        Flag to indicate whether GPU is used, by default False.
    """

    def __init__(
        self,
        list_correction_methods: list[tuple[LiteralModelCorrection, dict]],
        list_detection_methods: list[tuple[LiteralModelDetection, dict]],
        use_gpu: bool = False,
    ) -> None:
        """
        Initialize the ColorCorrectionAnalyzer.

        Parameters
        ----------
        list_correction_methods : list of tuple[LiteralModelCorrection, dict]]
            List of correction methods and their parameters.
        list_detection_methods : list of tuple[LiteralModelDetection, dict]]
            List of detection methods and their parameters.
        use_gpu : bool, optional
            Whether to use GPU acceleration, by default True.
        """
        self.list_correction_methods = list_correction_methods
        self.list_detection_methods = list_detection_methods
        self.use_gpu = use_gpu
        self.rg = ReportGenerator()

    def _run_single_exp(
        self,
        idx: int,
        input_image: np.ndarray,
        det_method: LiteralModelDetection,
        det_params: dict,
        cc_method: LiteralModelCorrection,
        cc_params: dict,
        reference_image: np.ndarray | None = None,
    ) -> dict:
        """
        Run a single experiment for a given detection and correction method.

        Parameters
        ----------
        idx : int
            Index of the experiment.
        input_image : np.ndarray
            The input image array.
        det_method : LiteralModelDetection
            The detection method identifier.
        det_params : dict
            Parameters for the detection method.
        cc_method : LiteralModelCorrection
            The correction method identifier.
        cc_params : dict
            Parameters for the correction method.
        reference_image : np.ndarray, optional
            The reference image, by default None.

        Returns
        -------
        dict
            A dictionary containing evaluation data and results of the experiment.
        """
        cc = ColorCorrection(
            correction_model=cc_method,
            detection_model=det_method,
            detection_conf_th=det_params.get("detection_conf_th", 0.25),
            use_gpu=self.use_gpu,
            **cc_params,
        )

        if reference_image is not None:
            cc.set_reference_image(reference_image)
        cc.set_input_patches(input_image, debug=True)
        cc.fit()
        corrected_image = cc.predict(input_image=input_image)
        eval_results = cc.calc_color_diff_patches()

        before_comparison = visualize_patch_comparison(
            ls_mean_in=cc.input_patches,
            ls_mean_ref=cc.reference_patches,
        )
        after_comparison = visualize_patch_comparison(
            ls_mean_in=cc.corrected_patches,
            ls_mean_ref=cc.reference_patches,
        )

        dE_image = calc_color_diff(  # noqa: N806
            image1=input_image,
            image2=corrected_image,
        )

        one_row = {
            "Index": idx,
            "Detection Method": det_method,
            "Detection Parameters": det_params,
            "Drawed Preprocessing Input": cc.input_debug_image,
            "Drawed Preprocessing Reference": cc.reference_debug_image,
            "Correction Method": cc_method,
            "Correction Parameters": cc_params,
            "Color Patches - Before": before_comparison,
            "Color Patches - After": after_comparison,
            "Input Image": input_image,
            "Corrected Image": corrected_image,
            "Patch ΔE (Before) - Min": eval_results["initial"]["min"],
            "Patch ΔE (Before) - Max": eval_results["initial"]["max"],
            "Patch ΔE (Before) - Mean": eval_results["initial"]["mean"],
            "Patch ΔE (Before) - Std": eval_results["initial"]["std"],
            "Patch ΔE (After) - Min": eval_results["corrected"]["min"],
            "Patch ΔE (After) - Max": eval_results["corrected"]["max"],
            "Patch ΔE (After) - Mean": eval_results["corrected"]["mean"],
            "Patch ΔE (After) - Std": eval_results["corrected"]["std"],
            "Image ΔE - Min": dE_image["min"],
            "Image ΔE - Max": dE_image["max"],
            "Image ΔE - Mean": dE_image["mean"],
            "Image ΔE - Std": dE_image["std"],
        }
        return one_row

    def run(
        self,
        input_image: np.ndarray,
        output_dir: str = "benchmark_debug",
        reference_image: np.ndarray | None = None,
    ) -> pd.DataFrame:
        """
        Run the full benchmark for color correction and generate reports.

        Parameters
        ----------
        input_image : np.ndarray
            The image to be processed.
        output_dir : str, optional
            The directory to save reports, by default `benchmark_debug`.
        reference_image : np.ndarray, optional
            Optional reference image used for evaluation, by default None.

        Returns
        -------
        pd.DataFrame
            A DataFrame containing results of all experiments.

        Notes
        -----
        The specified output directory will be created if it does not exist. All benchmark results are saved in this folder.
        The directory to save benchmark results, by default "benchmark_debug". This folder will contain:

        - **An HTML report**: includes a matrix table showing correction methods vs. evaluation delta E (CIE 2000) and preview images.
        - **A CSV file**: A CSV report of the DataFrame with image data columns removed.
        - **A PKL file**: A pickle file containing the full DataFrame.

        Examples
        --------
        >>> import numpy as np
        >>> from color_correction.services.correction_analyzer import ColorCorrectionAnalyzer
        >>>
        >>> # Assume input_image is a numpy array of shape (H, W, C)
        >>> input_image = np.random.rand(100, 100, 3)
        >>> analyzer = ColorCorrectionAnalyzer(
        ...     list_correction_methods=[
        ...         ("least_squares", {}),
        ...         ("linear_reg", {}),
        ...         ("affine_reg", {}),
        ...         ("polynomial", {"degree": 2})
        ...         ("polynomial", {"degree": 3})
        ...         ("polynomial", {"degree": 4})
        ...     ],
        ...     list_detection_methods=[("yolov8", {"detection_conf_th": 0.25})],
        ... )
        >>> results = analyzer.run(input_image=input_image, output_dir="output")
        >>> print(results.head())
        """  # noqa: E501
        ls_data = []
        idx = 1
        for det_method, det_params in self.list_detection_methods:
            for cc_method, cc_params in self.list_correction_methods:
                print(
                    f"Running benchmark for {cc_method} method with {cc_params}",
                )
                data = self._run_single_exp(
                    idx=idx,
                    input_image=input_image,
                    det_method=det_method,
                    det_params=det_params,
                    cc_method=cc_method,
                    cc_params=cc_params,
                    reference_image=reference_image,
                )
                idx += 1
                ls_data.append(data)
        df_results = pd.DataFrame(ls_data)

        # Generate HTML report path
        os.makedirs(output_dir, exist_ok=True)
        html_report_path = os.path.join(output_dir, "report.html")
        pickel_report_path = os.path.join(output_dir, "report.pkl")

        # Report Generator -----------------------------------------------------
        self.rg.generate_html_report(df=df_results, path_html=html_report_path)
        self.rg.save_dataframe(df=df_results, filepath=pickel_report_path)

        # Save CSV report, but without image data
        df_results.drop(
            columns=[
                "Drawed Preprocessing Input",
                "Drawed Preprocessing Reference",
                "Color Patches - Before",
                "Color Patches - After",
                "Corrected Image",
                "Input Image",
            ],
        ).to_csv(os.path.join(output_dir, "report_no_image.csv"), index=False)

        print(f"Report saved to {html_report_path}")
        return df_results

Functions

run 

run(input_image: np.ndarray, output_dir: str = 'benchmark_debug', reference_image: np.ndarray | None = None) -> pd.DataFrame

Run the full benchmark for color correction and generate reports.

Parameters:

Name Type Description Default
input_image np.ndarray

The image to be processed.

 required
output_dir str

The directory to save reports, by default benchmark_debug.

 'benchmark_debug'
reference_image np.ndarray

Optional reference image used for evaluation, by default None.

 None

Returns:

Type Description
pd.DataFrame

A DataFrame containing results of all experiments.
Notes

The specified output directory will be created if it does not exist. All benchmark results are saved in this folder. The directory to save benchmark results, by default "benchmark_debug". This folder will contain:

  • An HTML report: includes a matrix table showing correction methods vs. evaluation delta E (CIE 2000) and preview images.
  • A CSV file: A CSV report of the DataFrame with image data columns removed.
  • A PKL file: A pickle file containing the full DataFrame.

Examples:

>>> import numpy as np
>>> from color_correction.services.correction_analyzer import ColorCorrectionAnalyzer
>>>
>>> # Assume input_image is a numpy array of shape (H, W, C)
>>> input_image = np.random.rand(100, 100, 3)
>>> analyzer = ColorCorrectionAnalyzer(
...     list_correction_methods=[
...         ("least_squares", {}),
...         ("linear_reg", {}),
...         ("affine_reg", {}),
...         ("polynomial", {"degree": 2})
...         ("polynomial", {"degree": 3})
...         ("polynomial", {"degree": 4})
...     ],
...     list_detection_methods=[("yolov8", {"detection_conf_th": 0.25})],
... )
>>> results = analyzer.run(input_image=input_image, output_dir="output")
>>> print(results.head())
Source code in color_correction/services/correction_analyzer.py 
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
def run(
    self,
    input_image: np.ndarray,
    output_dir: str = "benchmark_debug",
    reference_image: np.ndarray | None = None,
) -> pd.DataFrame:
    """
    Run the full benchmark for color correction and generate reports.

    Parameters
    ----------
    input_image : np.ndarray
        The image to be processed.
    output_dir : str, optional
        The directory to save reports, by default `benchmark_debug`.
    reference_image : np.ndarray, optional
        Optional reference image used for evaluation, by default None.

    Returns
    -------
    pd.DataFrame
        A DataFrame containing results of all experiments.

    Notes
    -----
    The specified output directory will be created if it does not exist. All benchmark results are saved in this folder.
    The directory to save benchmark results, by default "benchmark_debug". This folder will contain:

    - **An HTML report**: includes a matrix table showing correction methods vs. evaluation delta E (CIE 2000) and preview images.
    - **A CSV file**: A CSV report of the DataFrame with image data columns removed.
    - **A PKL file**: A pickle file containing the full DataFrame.

    Examples
    --------
    >>> import numpy as np
    >>> from color_correction.services.correction_analyzer import ColorCorrectionAnalyzer
    >>>
    >>> # Assume input_image is a numpy array of shape (H, W, C)
    >>> input_image = np.random.rand(100, 100, 3)
    >>> analyzer = ColorCorrectionAnalyzer(
    ...     list_correction_methods=[
    ...         ("least_squares", {}),
    ...         ("linear_reg", {}),
    ...         ("affine_reg", {}),
    ...         ("polynomial", {"degree": 2})
    ...         ("polynomial", {"degree": 3})
    ...         ("polynomial", {"degree": 4})
    ...     ],
    ...     list_detection_methods=[("yolov8", {"detection_conf_th": 0.25})],
    ... )
    >>> results = analyzer.run(input_image=input_image, output_dir="output")
    >>> print(results.head())
    """  # noqa: E501
    ls_data = []
    idx = 1
    for det_method, det_params in self.list_detection_methods:
        for cc_method, cc_params in self.list_correction_methods:
            print(
                f"Running benchmark for {cc_method} method with {cc_params}",
            )
            data = self._run_single_exp(
                idx=idx,
                input_image=input_image,
                det_method=det_method,
                det_params=det_params,
                cc_method=cc_method,
                cc_params=cc_params,
                reference_image=reference_image,
            )
            idx += 1
            ls_data.append(data)
    df_results = pd.DataFrame(ls_data)

    # Generate HTML report path
    os.makedirs(output_dir, exist_ok=True)
    html_report_path = os.path.join(output_dir, "report.html")
    pickel_report_path = os.path.join(output_dir, "report.pkl")

    # Report Generator -----------------------------------------------------
    self.rg.generate_html_report(df=df_results, path_html=html_report_path)
    self.rg.save_dataframe(df=df_results, filepath=pickel_report_path)

    # Save CSV report, but without image data
    df_results.drop(
        columns=[
            "Drawed Preprocessing Input",
            "Drawed Preprocessing Reference",
            "Color Patches - Before",
            "Color Patches - After",
            "Corrected Image",
            "Input Image",
        ],
    ).to_csv(os.path.join(output_dir, "report_no_image.csv"), index=False)

    print(f"Report saved to {html_report_path}")
    return df_results