Crops image according to mod to restore spatial dimensions adequately in the decoding sections of the model. :param image: numpy array

Create image ensemble to estimate denoised image. :param image: numpy array Noisy image. :param normalize: bool Normalize

Training of the denoiser model. :param model: torch Module Neural network to fit. :param data_loaders: dict Dictionary wi

Returns the average of the monitored metrics in the current moment, given the number of evaluated samples. :param n_samples:

Load image and change it color space from RGB to Grayscale if necessary. :param image_path: str Path of the image. :param channel

Pads image according to mod to restore spatial dimensions adequately in the decoding sections of the model. :param image: numpy array

Predict batch of images from an ensemble. :param model: torch Module Trained model to estimate denoised images. :param ensemble:

Apply inverse transforms to predicted image ensemble and average them. :param ensemble: list Predicted images, ensemble[0] is the ori

Sets all random seeds. :param seed: int Seed value. :return: None

Update the metrics in the current epoch, this method is called at every step of the epoch. :param metrics: dict Metrics t

Updates the CSV record file. :param epoch_log: dict Log of the current epoch. :return: None

Generates additive white gaussian noise, and it is applied to the clean image. :param sample: :return:

Creates the csv record file. :param f