swirlspy.qpf.dl.helpers package

Submodules

swirlspy.qpf.dl.helpers.gifmaker module

swirlspy.qpf.dl.helpers.gifmaker.save_gif(single_seq, fname)

Save a single gif consisting of image sequence in single_seq to fname.

swirlspy.qpf.dl.helpers.gifmaker.save_gifs(seq, prefix)

Save several gifs.

Parameters

  • seq (Shape (num_gifs, IMG_SIZE, IMG_SIZE)) –

  • prefix (prefix-idx.gif will be the final filename.) –

swirlspy.qpf.dl.helpers.log_analysis module

swirlspy.qpf.dl.helpers.log_analysis.parse_log(file_path, regex)
Parameters

  • file_path –

  • regex –

swirlspy.qpf.dl.helpers.log_analysis.remove_duplicates_and_convert_npy(val_list)
swirlspy.qpf.dl.helpers.log_analysis.temporal_smoothing(training_statistics, stride=10, window_size=100)

We always assume the first axis in statistics is the iteration

Parameters

  • training_statistics –

  • stride –

  • window_size –

Returns

  • smoothed_mean

  • smoothed_std

swirlspy.qpf.dl.helpers.msssim module

Python implementation of MS-SSIM. Usage: python msssim.py –original_image=original.png –compared_image=distorted.png

swirlspy.qpf.dl.helpers.msssim.MultiScaleSSIM(img1, img2, max_val=255, filter_size=11, filter_sigma=1.5, k1=0.01, k2=0.03, weights=None)

Return the MS-SSIM score between img1 and img2. This function implements Multi-Scale Structural Similarity (MS-SSIM) Image Quality Assessment according to Zhou Wang’s paper, “Multi-scale structural similarity for image quality assessment” (2003). Link: https://ece.uwaterloo.ca/~z70wang/publications/msssim.pdf Author’s MATLAB implementation: http://www.cns.nyu.edu/~lcv/ssim/msssim.zip

Parameters

  • img1 (Numpy array holding the first RGB image batch.) –

  • img2 (Numpy array holding the second RGB image batch.) –

  • max_val (the dynamic range of the images (i.e., the difference between the) – maximum the and minimum allowed values).

  • filter_size (Size of blur kernel to use (will be reduced for small images).) –

  • filter_sigma (Standard deviation for Gaussian blur kernel (will be reduced) – for small images).

  • k1 (Constant used to maintain stability in the SSIM calculation (0.01 in) – the original paper).

  • k2 (Constant used to maintain stability in the SSIM calculation (0.03 in) – the original paper).

  • weights (List of weights for each level; if none, use five levels and the) – weights from the original paper.

Return type

MS-SSIM score between img1 and img2.

Raises

RuntimeError – If input images don’t have the same shape or don’t have four: dimensions: [batch_size, height, width, depth].

swirlspy.qpf.dl.helpers.ordered_easydict module

class swirlspy.qpf.dl.helpers.ordered_easydict.OrderedEasyDict(d=None, **kwargs)

Bases: OrderedDict

Using OrderedDict for the easydict package See Also https://pypi.python.org/pypi/easydict/

swirlspy.qpf.dl.helpers.visualization module

swirlspy.qpf.dl.helpers.visualization.flow_to_img(flow_dat, max_displacement=None)

Convert optical flow data to HSV images

Parameters

  • flow_dat (np.ndarray) – Shape: (seq_len, 2, H, W)

  • max_displacement (float or None) –

Returns

rgb_dat – Shape: (seq_len, 3, H, W)

Return type

np.ndarray

swirlspy.qpf.dl.helpers.visualization.get_color_flow_legend_image(size=50)
swirlspy.qpf.dl.helpers.visualization.merge_rgba_cv2(front_img, back_img)

Merge the front image with the background image using the Painter’s algorithm

Parameters

  • front_img (np.ndarray) –

  • back_img (np.ndarray) –

Returns

result_img

Return type

np.ndarray

swirlspy.qpf.dl.helpers.visualization.save_hko_gif(im_dat, save_path)

Save the HKO images to gif

Parameters

  • im_dat (np.ndarray) – Shape: (seqlen, H, W)

  • save_path (str) –

swirlspy.qpf.dl.helpers.visualization.save_hko_movie(im_dat, datetime_list, mask_dat=None, save_path='hko.mp4', masked=False, fps=5, prediction_start=None)

Save the HKO images to a video file

Parameters

  • im_dat (np.ndarray) – Shape : (seq_len, H, W)

  • datetime_list (list) – list of datetimes

  • mask_dat (np.ndarray or None) – Shape : (seq_len, H, W)

  • save_path (str) –

  • masked (bool) – whether the mask the inputs when saving the image

  • fps (float) – the fps of the saved movie

  • prediction_start (int or None) – The starting point of the prediction