Data and Scripts
for Hydrological Streamline Detection Using a U-net Model with Attention Module
$Zewei$ $Xu^{1,2}$; $Shaowen$ $Wang^{1,2}$; $Lawrence V.$ $Stanislawski^{3}$; $Zhe$ $Jiang^{4}$; $Nattapon$ $Jaroenchai^{1,2}$; $Arpan Man$ $Sainju^{4}$; $Ethan$ $Shavers^{3}$; $E. Lynn$ $Usery^{3}$; $Li$ $Chen^{2,5}$; $Zhiyu$ $Li^{1,2}$; $Bin$ $Su^{1,2}$
$^{1}$$Department$ $of$ $Geography$ $and$ $Geographic$ $Information$ $Science$, $University$ $of$ $Illinois$ $at$ $Urbana-Champaign$, $Urbana$, $IL$, $USA$
$^{2}$$CyberGIS$ $Center$ $for$ $Advanced$ $Digital$ $and$ $Spatial$ $Studies$, $University$ $of$ $Illinois$ $at$ $Urbana-Champaign$, $Urbana$, $IL$, $USA$
$^{3}$$U.S.$ $Geological$ $Survey$, $Center$ $of$ $Excellence$ $for$ $Geospatial$ $Information$ $Science$, $Rolla$, $MO$, $USA$
$^{4}$$Department$ $of$ $Computer$ $Science$, $University$ $of$ $Alabama$, $Tuscaloosa$, $AL$, $USA$
$^{5}$$School$ $of$ $Geosciences$ $and$ $Info-Physics$, $Central$ $South$ $University$, $Changsha$, $Hunan$, $China$
$Corresponding$ $Author:$ $nj7@illinois.edu$
Notebook Structure:
1. Introduction
Interpret the Result¶
This notebook contains the code that we used to extract statistics and evaluate the result of CNN training.
The cell below is to initiate the function that will be used to extract trainging and validating statistics during the CNN training.
root="/home/jovyan/shared_data/data/unet_streamline_detection/"
m='v2'
aug = '_aug'+m #+sys.argv[1] #'v2'
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import itertools
import os
import pickle
from itertools import chain
from skimage.io import imread, imshow, imread_collection, concatenate_images
from skimage.transform import resize
from skimage.morphology import label
from sklearn.metrics import confusion_matrix
"""
plot_history(): reads Keras result and ogenerate figure of Loss, Dice Coefficient, and Accuracy.
"""
def plot_history(history):
loss_list = [s for s in history.keys() if 'loss' in s and 'val' not in s]
val_loss_list = [s for s in history.keys() if 'loss' in s and 'val' in s]
dice_list = [s for s in history.keys() if 'dice' in s and 'val' not in s]
val_dice_list = [s for s in history.keys() if 'dice' in s and 'val' in s]
acc_list = [s for s in history.keys() if 'acc' in s and 'val' not in s]
val_acc_list = [s for s in history.keys() if 'acc' in s and 'val' in s]
if len(loss_list) == 0:
print('Loss is missing in history')
return
## As loss always exists
epochs = range(1,len(history[loss_list[0]]) + 1)
## Loss
plt.subplot(1,3,1)
for l in loss_list:
plt.plot(epochs, history[l], 'b', label='Training loss (' + str(str(format(history[l][-1],'.5f'))+')'))
for l in val_loss_list:
plt.plot(epochs, history[l], 'g', label='Validation loss (' + str(str(format(history[l][-1],'.5f'))+')'))
plt.title('Loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend()
## Dice Coefficient
plt.subplot(1,3,2)
for l in dice_list:
plt.plot(epochs, history[l], 'b', label='Training Dice Coefficient (' + str(format(history[l][-1],'.5f'))+')')
for l in val_dice_list:
plt.plot(epochs, history[l], 'g', label='Validation Dice Coefficient (' + str(format(history[l][-1],'.5f'))+')')
plt.title('Dice Coefficient')
plt.xlabel('Epochs')
plt.ylabel('Dice Coefficient')
plt.legend()
## Accuracy
plt.subplot(1,3,3)
for l in acc_list:
plt.plot(epochs, history[l], 'b', label='Training accuracy (' + str(format(history[l][-1],'.5f'))+')')
for l in val_acc_list:
plt.plot(epochs, history[l], 'g', label='Validation accuracy (' + str(format(history[l][-1],'.5f'))+')')
plt.title('Accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.legend()
plt.show()
Plot the Result¶
The plot is generated based on the log of the trainging and validating processes which contains the statistic of training and validating phases.
infile = open(root+'result/history_augn2_attention2.pickle','rb')
new_dict = pickle.load(infile)
infile.close()
plt.figure(figsize=(20,5))
plot_history(new_dict)
import copy
import random
import sys
import numpy as np
buf =30
"""
plot_confusion_matrix(): prints and plots the confusion matrix. Normalization can be applied by setting `normalize=True`.
"""
def plot_confusion_matrix(cm, classes,
normalize=False,
title='Confusion matrix',
cmap=plt.cm.Blues):
if normalize:
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
print("Normalized confusion matrix")
else:
print('Confusion matrix, without normalization')
print(cm)
plt.imshow(cm, interpolation='nearest', cmap=cmap)
plt.title(title)
plt.colorbar()
tick_marks = np.arange(len(classes))
plt.xticks(tick_marks, classes, rotation=45)
plt.yticks(tick_marks, classes)
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):
plt.text(j, i, format(cm[i, j], fmt),
horizontalalignment="center",
color="white" if cm[i, j] > thresh else "black")
plt.tight_layout()
plt.ylabel('True label')
plt.xlabel('Predicted label')
preds_test_mod = np.load(root_path+'result/preds_test_total_augn2_attention2.npy')
dim = np.load(root+'data/reference.npy').shape
numr = dim[0]//(224 - buf*2)
numc = dim[1]//(224 - buf*2)
count = -1
for i in range(numr):
for j in range(numc):
count += 1
temp = preds_test_mod[count][buf:-buf,buf:-buf]
if j == 0:
rows = temp
else:
rows = np.concatenate((rows,temp),axis = 1)
if i == 0:
prediction_map = copy.copy(rows)
else:
prediction_map = np.concatenate((prediction_map,rows),axis = 0)
prediction_map = prediction_map[:,:,0]
# fig= plt.figure(figsize=(20,15))
# plt.imshow(prediction_map*255,cmap='gray',vmin=0, vmax=255)
# plt.show()
#print(np.unique(prediction_map))
# mask
mask = np.load(root+'data/mask.npy')[:prediction_map.shape[0],:prediction_map.shape[1]]
[lr,lc] = np.where(mask == 1)
# Read reference data
groundtruthlist = np.load(root+'data/reference.npy')[:prediction_map.shape[0],:prediction_map.shape[1]][lr,lc]
predictionlist = prediction_map[lr,lc]
# print(np.unique(groundtruthlist*255))
# print(np.unique(predictionlist))
# print(type(groundtruthlist))
# print(groundtruthlist)
# print(type(predictionlist))
# print(predictionlist)
cm = confusion_matrix((groundtruthlist).astype(int), predictionlist)
plot_confusion_matrix(cm,classes=["Non-streams","Streams"])
The cell below will calculate 3 statistics that can be used to evaluate the performance of the model.
# Statistics
from sklearn.metrics import f1_score, precision_score,recall_score
#print(f1_score(groundtruthlist, predictionlist, average='macro'))
print('F1 score of stream: '+str(f1_score(groundtruthlist, predictionlist,pos_label=1)))
print('F1 score of nonstream: '+str(f1_score(groundtruthlist, predictionlist,pos_label=0)))
print('Precision of stream: '+str(precision_score(groundtruthlist, predictionlist,pos_label=1)))
print('Precision of nonstream: '+str(precision_score(groundtruthlist, predictionlist,pos_label=0)))
print('Recall of stream: '+str(recall_score(groundtruthlist, predictionlist,pos_label=1)))
print('Recall of nonstream: '+str(recall_score(groundtruthlist, predictionlist,pos_label=0)))
The cell below is the code that show the result compare to the reference dataset
# Generate prediction map
from osgeo import gdal
import numpy as np
import copy
import matplotlib.pyplot as plt
fig= plt.figure(figsize=(20,15))
# Plot the map
prediction_map = prediction_map*mask
plt.subplot(1,2, 1)
plt.title('result',fontsize = 16)
plt.imshow(prediction_map*255,cmap='gray',vmin=0, vmax=1)
reference = np.load(root+'data/reference.npy')
plt.subplot(1,2,2)
plt.title('Reference map',fontsize = 16)
plt.imshow(reference*255,cmap='gray',vmin=0, vmax=255)
plt.show()
