Companies fulfill the need for data by using the\u00a0data augmentation<\/strong>\u00a0technique that reduces the work of collecting and preparing training data beforehand supplying it to the model.<\/p>\n\n\n\n In this article, you’ll be introduced to –<\/p>\n\n\n\n Data augmentation<\/strong> is a process of increasing the quantity of data artificially by generating new data points from the existing data. This includes adding minor alterations to the data or using machine learning models to generate new data points.<\/p>\n\n\n\n If you are working with image data, then augmenting that data would be flipping the images horizontally and vertically, increasing or decreasing the brightness, skewing, rotating, shifting the position, and more.<\/p>\n\n\n\n In production use cases, collecting the training data can be costly and highly time-consuming. However, the company uses data augmentation, an effective and low-cost method that cuts off the work of collection and preparation of new data.<\/p>\n\n\n\n Data augmentation is useful for improving the performance of the machine learning model which will eventually affect the outcome in a good way by diversifying the dataset.<\/p>\n\n\n\n The machine learning model’s accuracy completely depends on the dataset’s quality, quantity and representability. Collecting and cleaning the new data may increase the risk of the existing data being less accurate and representable. Data augmentation techniques help produce new variations from the current data that make the machine-learning model more robust and perform well for real-world inputs.<\/p>\n\n\n\n Data augmentation is about making minor changes to the existing data to produce new variations. These are done by passing the dataset to transformation functions that perform transformation and then produce a new dataset.<\/p>\n\n\n\n Here’s an illustration that will explain the workflow of the data augmentation process in which a sequence of transformation functions are applied by human experts.<\/p>\n\n\n\n A usual workflow of heuristic data augmentation has the following steps:<\/p>\n\n\n\n a. The data is fed to the data augmentation pipeline containing the sequence of transformation functions.<\/p>\n\n\n\n b. The data augmentation pipeline is defined by the sequential steps of different augmentations to produce different variations of each data point.<\/p>\n\n\n\n c. The image is then fed to the data augmentation pipeline and processed through each step of the transformation function.<\/p>\n\n\n\n d. After the image is processed, a human expert verifies the augmented results.<\/p>\n\n\n\n e. After the verification process, the augmented results are now ready to use by the AI model for training.<\/p>\n\n\n\n Heuristic data augmentation is used to work with models that will classify images<\/strong>.<\/p>\n<\/blockquote>\n\n\n\n Data augmentation is less prevalent in the NLP<\/strong> (Natural Language Processing) field than in image data. The main reason is the complexity of the natural language makes it difficult to automate the process of augmenting the text data.<\/p>\n\n\n\n Some popular open-source Python packages for data augmentation available are ImageDataGenerator from Keras, Skimage, and OpenCV.<\/p>\n\n\n\n But there are also some advanced models for data augmentation available that gained popularity lately.<\/p>\n\n\n\n For Natural Language Processing<\/strong><\/p>\n\n\n\n For demonstration purposes, you’re going to see the data augmentation techniques for the image using a deep-learning library called Keras<\/strong> that provides a utility called ImageDataGenerator<\/strong>.<\/p>\n\n\n\n Let’s start coding the process of augmenting the image data and here is the preview of the sample image that will be used.<\/p>\n\n\n\n Image Credit<\/strong>: Photo by Juan Manuel Sanchez<\/a> on Unsplash<\/a>.<\/p>\n<\/blockquote>\n\n\n\n The augmentation techniques you’re going to see are:<\/p>\n\n\n\n Visit my GitHub<\/a> repository to learn additional data-augmenting techniques.<\/p>\n\n\n\n In the following notebook, we are randomly shifting the image by width and height.<\/p>\n\n\n\n First, we imported the required libraries and loaded our image and then we converted the image into an array and expanded the dimension of the array.<\/p>\n\n\n\n Next, we instantiated the To see the augmented results, we iterated over the original image, converted it into an unsigned integer, and then plotted the augmented samples.<\/p>\n\n\n\n We are doing the same process but this time we changed the augmenting technique and specified the operation To generate the rotated images, we specified Here, we are randomly increasing and decreasing the brightness of the image by providing Data augmentation is useful when there is less amount of data or is often used for reducing the load of collecting and preparing new data. Many companies see it as a low-cost and effective method to train their AI models.<\/p>\n\n\n\n Data augmentation is more popular with image data than in the NLP domain. The complexity of natural language makes it harder to automate the augmentation of text data.<\/p>\n\n\n\n Let’s point out the topics you’ve learned in this article:<\/p>\n\n\n\n With all that, you’ve seen the example of performing the different augmenting techniques on image data like<\/p>\n\n\n\n There are only four examples shown but if you want to see more augmenting techniques, visit GitHub<\/a> for source code.<\/p>\n\n\n\n \ud83c\udfc6Other articles you might like<\/strong><\/p>\n\n\n\n \u2705How to implement a custom deep learning model into the Flask app for image recognition<\/a>.<\/p>\n\n\n\n \u2705Numpy and TensorFlow have a common function<\/a>.<\/p>\n\n\n\n \u2705Merge the TailwindCSS into Flask apps to make more interactive webpages<\/a>.<\/p>\n\n\n\n \u2705Take multiple inputs from the user in a single command<\/a>.<\/p>\n\n\n\n \u2705Perform high-level file operations using the shutil module<\/a>.<\/p>\n\n\n\n\n
<\/a>What is data augmentation?<\/h1>\n\n\n\n
<\/a>Why do we need data augmentation?<\/h1>\n\n\n\n
<\/a>How does data augmentation work?<\/h1>\n\n\n\n
![\"Figure:](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/Data-1024x448.png\")
<\/figure>\n\n\n\n\n
\n
<\/a>Models for data augmentation<\/h1>\n\n\n\n
\n
\n
<\/a>Demonstration of data augmentation<\/h1>\n\n\n\n
![\"Sample](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/cat.png\")
<\/figure>\n\n\n\n\n
\n
<\/a>Random Shift<\/h3>\n\n\n\n
import numpy as np\nimport matplotlib.pyplot as plt\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.utils import load_img, img_to_array\n\n# Preview of the image we will work on\nimg = load_img('\/content\/drive\/MyDrive\/Sample\/resized_cat.jpg')\n\n# Random Shift\npic = img_to_array(img)\npic_sample = np.expand_dims(pic, axis=0)\n\ndata = ImageDataGenerator(width_shift_range=[10, 50], height_shift_range=[30, 70])\naug_data = data.flow(pic_sample, batch_size=1)\n\nfig, axes = plt.subplots(1, 10, figsize=(30, 40))\n\nfor i in range(10):\n # Accessing each sample\n aug_img = next(aug_data)\n # Converting to unsigned integer\n aug_result = aug_img[0].astype('uint8')\n # Showing augmented images\n axes[i].imshow(aug_result)\n axes[i].axis('off')<\/pre><\/div>\n\n\n\n![\"\"](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/random_shift-1024x117.png\")
<\/figure>\n\n\n\nImageDataGenerator<\/code> class and specified the operation width_shift_range=[10, 50], height_shift_range=[30, 70]<\/code> and then passed the sample image to the ImageDataGenerator<\/em><\/strong> for generating the processed images.<\/p>\n\n\n\n<\/a>Random Flip<\/h3>\n\n\n\n
horizontal_flip=True, vertical_flip=True<\/code> that will randomly flip the images horizontally and vertically.<\/p>\n\n\n\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.utils import load_img, img_to_array\n\n# Preview of the image we will work on\nimg = load_img('\/content\/drive\/MyDrive\/Sample\/resized_cat.jpg')\n\n# Random flip\npic = img_to_array(img)\npic_sample = np.expand_dims(pic, axis=0)\n\ndata = ImageDataGenerator(horizontal_flip=True, vertical_flip=True)\naug_data = data.flow(pic_sample, batch_size=1)\n\nfig, axes = plt.subplots(1, 10, figsize=(30, 40))\n\nfor i in range(10):\n # Accessing each sample\n aug_img = next(aug_data)\n # Converting to unsigned integer\n aug_result = aug_img[0].astype('uint8')\n # Showing augmented images\n axes[i].imshow(aug_result)\n axes[i].axis('off')<\/pre><\/div>\n\n\n\n![\"\"](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/random_flip-1024x117.png\")
<\/figure>\n\n\n\nRotation<\/h3>\n\n\n\n
rotation_range=50<\/code> that will randomly generate the images within the specified range.<\/p>\n\n\n\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.utils import load_img, img_to_array\n\n# Preview of the image we will work on\nimg = load_img('\/content\/drive\/MyDrive\/Sample\/resized_cat.jpg')\n\n# Rotation\npic = img_to_array(img)\npic_sample = np.expand_dims(pic, axis=0)\n\ndata = ImageDataGenerator(rotation_range=50)\naug_data = data.flow(pic_sample, batch_size=1)\n\nfig, axes = plt.subplots(1, 5, figsize=(30, 40))\n\nfor i in range(5):\n # Accessing each sample\n aug_img = next(aug_data)\n # Converting to unsigned integer\n aug_result = aug_img[0].astype('uint8')\n # Showing augmented images\n axes[i].imshow(aug_result)\n axes[i].axis('off')<\/pre><\/div>\n\n\n\n![\"\"](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/rotation-1024x230.png\")
<\/figure>\n\n\n\nRandom Brightness<\/h3>\n\n\n\n
brightness_range=[0.5, 1.5]<\/code>.<\/p>\n\n\n\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.utils import load_img, img_to_array\n\n# Preview of the image we will work on\nimg = load_img('\/content\/drive\/MyDrive\/Sample\/resized_cat.jpg')\n\n# Random Brightness\npic = img_to_array(img)\npic_sample = np.expand_dims(pic, axis=0)\n\ndata = ImageDataGenerator(brightness_range=[0.5, 1.5])\naug_data = data.flow(pic_sample, batch_size=1)\n\nfig, axes = plt.subplots(1, 5, figsize=(30, 40))\n\nfor i in range(5):\n # Accessing each sample\n aug_img = next(aug_data)\n # Converting to unsigned integer\n aug_result = aug_img[0].astype('uint8')\n # Showing augmented images\n axes[i].imshow(aug_result)\n axes[i].axis('off')<\/pre><\/div>\n\n\n\n![\"\"](\"https:\/\/geekpython.in\/wp-content\/uploads\/2023\/08\/random_bright-1024x229.png\")
<\/figure>\n\n\n\nConclusion<\/h1>\n\n\n\n
\n
\n
\n\n\n\n