Scripts and functions to assist with common Visual AI tasks
| Script | Description |
|---|---|
| coco_to_pascal.py | Convert COCO-style annotation files (.json) into Pascal VOC format annotation files (.xml). |
| yolo_to_pascal.py | Convert YOLO format annotation files (.txt) to Pascal VOC format annotation files (.xml). |
| process_coco_for_segmentation.py | Process COCO annotations, modify category IDs, generate a defect map, and create segmentation masks for each image. |
| transparent_background_to_solid.py | Convert all images in a directory with transparency to RGB mode, and save them with a specified background color. (e.g. white) |
| extract_images_from_videos.py | Extract frames from video files in a directory at a specified frame rate and save them as images. |
| convert_jfif_to_jpeg.py | Convert all .jfif files in a directory to .jpeg format, removing any alpha channel (transparency) if present. |
| find_def_location_grid.py | Determine the grid cell location of a bounding box within an image. |
| convert_png_masks.py | Convert all masks in in a directory to the LandingAI-compatible format for masks (8-bit grayscale). |
| crop_images_json.py | Crop images based on cropping information in JSON files and save cropped images in label-specific directories. |
| resize_images_and_labels.py | Resize and save corresponding image and label (mask) files with matching filenames. |