Open an image, drag/resize colored boxes, pick class from a dropdown, add multiple boxes, crop/rotate, swipe prev/next. All coords saved to one labels.json. Fast, no GPU.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Open an image inside VS Code, drag/resize the green box right on the lesion (or any object), add a second box for multi-object images, swipe through a folder, and save — coordinates are written straight to the YOLO sidecar .txt. No GPU, no server, no upload.
Features
View any image with its YOLO box drawn in green.
Drag inside to move, drag a corner to resize.
Add multiple boxes per image (A), delete the selected box (Del).
Change class with number keys 0-9 (names read from classes.txt).
Swipe prev/next image (Alt+← / Alt+→) — auto-saves before moving.
Saves to the sidecar .txt in YOLO format (cls cx cy w h, normalized).
Expected layout
dataset/
classes.txt # one class name per line
img001.jpg
img001.txt # YOLO: "cls cx cy w h" (one line per box)
img002.jpg
img002.txt
classes.txt is found by walking up parent folders, so per-class subfolders work too.
Usage
Open the folder of images in VS Code.
Right-click an image → Reopen Editor With… → SwiftBox Editor (or set priority to default in package.json to open on click).
