This is a small tool for annotating objects in images and saving bounding boxes as CSV file and optionally the image patches as PNG. Useful for extracting train/test data from visual data to use for machine learning and computer vision applications.

• OpenCV 3.0.0
• Boost 1.56.0

• --inputdir [path with JPG/PNG/BMP files]
• --csvfile [path to CSV for saving annotations]
• --outputdir [optional output path for saving patches]

• Click left mouse button to add annotation
• Click again to increase size of annotation (also +/- keys can be used)
• Press BACKSPACE to undo last annotation
• Press U to remove all annotations in current image
• Press SPACE to save all annotations and continue to next image
• Press ESC to close the application