Synthetic datasets

This brief presentation shows a few sample images from generated synthetic data focusing on Atomic Pattern Dictionary and sprse encoding

%matplotlib inline
%load_ext autoreload
%autoreload 2
import os, sys, glob
import pandas
import numpy as np
from skimage import io
import matplotlib.pylab as plt
sys.path += [os.path.abspath('.'), os.path.abspath('..')]  # Add path to root
import bpdl.data_utils as tl_utils

DEFAULT_PATH = '/mnt/30C0201EC01FE8BC/TEMP'
DEFAULT_DATASET = 'atomicPatternDictionary_v0'
DEFAULT_IMG_POSIX = '.png'
pathDataset = os.path.join(DEFAULT_PATH, DEFAULT_DATASET)
print ([os.path.basename(p) for p in glob.glob(os.path.join(pathDataset, '*'))])

['combination.csv', 'datasetBinary_defNoise', 'datasetBinary_deform', 'datasetBinary_noise', 'datasetBinary_raw', 'datasetFuzzy_defNoise', 'datasetFuzzy_deform', 'datasetFuzzy_noise', 'datasetFuzzy_raw', 'datasetFuzzy_raw_gauss-0.001', 'datasetFuzzy_raw_gauss-0.005', 'datasetFuzzy_raw_gauss-0.010', 'datasetFuzzy_raw_gauss-0.025', 'datasetFuzzy_raw_gauss-0.050', 'datasetFuzzy_raw_gauss-0.075', 'datasetFuzzy_raw_gauss-0.100', 'datasetFuzzy_raw_gauss-0.125', 'datasetFuzzy_raw_gauss-0.150', 'datasetFuzzy_raw_gauss-0.200', 'dictionary']

def showDatasetImages(pathBase, dataset, imPattern='*', nbSamples=None, perRow=5):
    imgs, names = tl_utils.dataset_load_images(os.path.join(pathBase, dataset), imPattern, nbSamples)
    nbRows = int(np.ceil(float(len(imgs)) / perRow))
    plt.figure(figsize=(10, nbRows * perRow))
    for i, img in enumerate(imgs):
        plt.subplot(nbRows, perRow, i + 1), plt.imshow(img, cmap=plt.cm.gray), plt.axis('off')
    plt.tight_layout()

showDatasetImages(pathDataset, 'datasetBinary_raw', nbSamples=5)

showDatasetImages(pathDataset, 'datasetFuzzy_raw', nbSamples=5)

showDatasetImages(pathDataset, 'datasetBinary_defNoise', nbSamples=5)

showDatasetImages(pathDataset, 'datasetFuzzy_defNoise', nbSamples=5)

showDatasetImages(pathDataset, 'dictionary', imPattern='pattern_*', nbSamples=5)

pCoding = os.path.join(pathDataset, tl_utils.CSV_NAME_WEIGHTS) # 'combination.csv'
df = pandas.read_csv(pCoding, index_col=0)
print (df.head())

   combination          name
0  0;0;0;1;0;0  sample_00000
1  1;0;0;0;0;1  sample_00001
2  1;0;0;0;0;0  sample_00002
3  0;0;0;0;1;0  sample_00003
4  0;0;0;0;0;1  sample_00004

/usr/local/lib/python3.5/dist-packages/IPython/kernel/__main__.py:2: FutureWarning: from_csv is deprecated. Please use read_csv(...) instead. Note that some of the default arguments are different, so please refer to the documentation for from_csv when changing your function calls
  from IPython.kernel.zmq import kernelapp as app