NeuAug: Machine Learning Augmentation

Ever feel that your dataset is too small? Use neuaug to generate some augmented data for your image processing or natural language processing needs

Requirements

• Python 3.6+
• PIL
• googletrans

Usage

Image Augmentation

Run the image augmentation script from the command line with the following syntax:-

python imgaug.py <folder of jpg images> <output folder> <number of augmented copies>

Example:

python imgaug.py input_images augmented_images 5

This command will:

1. Read all JPG images from the input_images folder
2. Create 5 augmented copies of each image
3. Save the augmented images in the augmented_images folder

Word Augmentation

Run the word augmentation script from the command line with the following syntax:-

python textaug.py <input .txt file> <output .txt file> <number of augmented words for each word>

Example:

python textaug.py input_words.txt augmented_words.txt 5

This command will:

1. Read words from input_words.txt
2. Create 5 augmented versions of each word
3. Save the original and augmented words in augmented_words.txt

Speech Augmentation

Run the speech augmentation script from the command line with the following syntax:-

python speechaug.py <folder of audio files> <output folder> <number of augmented copies>

Example:

python speechaug.py input_audio augmented_audio 5

This command will:

1. Read all audio files from the input_audio folder
2. Create 5 augmented copies of each audio file
3. Save the augmented audio files in the augemented_audio folder