How do I input a set of images from a directory into python to use as a training set?

I've been able to extract URL datasets and links to be able to be used as a training/testing dataset, however I want to expand this into images. Basically, if I have 150 images of cats, how would I be able to input this in and classify with it?

Current code that extracts from URL using IRIS dataset

import pandas
from pandas.plotting import scatter_matrix
import matplotlib.pyplot as plt
from sklearn import model_selection
from sklearn.metrics import classification_report
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.neighbors import KNeighborsClassifier
url = "https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data"
names = ['sepal-length', 'sepal-width', 'petal-length', 'petal-width', 'class']
dataset = pandas.read_csv(url, names=names)
print(dataset.shape)
print(dataset.head(20))
print(dataset.loc[1])
print(dataset.describe())
print(dataset.loc[1][0])
plt.show()
dataset.hist()
plt.show()
scatter_matrix(dataset)
plt.show()

array = dataset.values
X = array[:,0:4]
Y = array[:,4]
validation_size = 0.20
seed = 7

X_train, X_validation, Y_train, Y_validation = model_selection.train_test_split(X, Y, test_size=validation_size, random_state=seed)
seed = 7
scoring = 'accuracy'
models = []
models.append(('KNN', KNeighborsClassifier()))
# evaluate each model in turn
results = []
names = []
for name, model in models:
    kfold = model_selection.KFold(n_splits=10, random_state=seed)
    cv_results = model_selection.cross_val_score(model, X_train, Y_train, cv=kfold, scoring=scoring)
    results.append(cv_results)
    names.append(name)
    msg = "%s: %f (%f)" % (name, cv_results.mean(), cv_results.std())
    print(msg)


fig = plt.figure()
fig.suptitle('Algorithm Comparison')
ax = fig.add_subplot(111)
plt.boxplot(results)
ax.set_xticklabels(names)
plt.show()
knn = KNeighborsClassifier()
knn.fit(X_train, Y_train)
predictions = knn.predict(X_validation)
print(accuracy_score(Y_validation, predictions))
print(confusion_matrix(Y_validation, predictions))
print(classification_report(Y_validation, predictions))

2 answers

  • answered 2018-08-14 09:49 kevinkayaks

    You can use your library of choice to read images with sequential filenames

    import skimage as ski
    filenames = ['image-%03d.jpg'%n for n in range(150)]
    images = []
    for f in filenames:
        im = ski.imread(f)
        images.append(im)
    

    Then images is a list of images.

    You can also iterate through any sort of filenames, or pull only files from a directory with a certain extension using the os module. The principle is the same. Just construct filenames as necessary.

    However, I recommend using pims, possibly with a processing pipeline

    import pims
    import numpy as np
    images = pims.ImageSequence('images-*.jpg')
    
    @pims.pipeline
    def grayarr(im):
        return np.array(im)[:,:,0]
    
    images = grayarr(images)
    

    At this point you can index into images with numpy-like slicing. pims is especially helpful when you're dealing with so many images you can't hold them in RAM. You can read about these things in the pims documentation.

  • answered 2018-08-14 11:45 Silas Jojo

    You could use Glob and extract the data from directory

    from PIL import Image
    import glob
    list_of_images = []
    
    for filename in glob.glob('file_directory/.jpg'): #assuming you are dealing with #jpg
        training_set = Image.open(filename)
        list_of_images.append(training_set)