Image-Augmentation

Why we included this project

Small labeled datasets are the usual bottleneck when you train a detector or segmentation model, and this tool attacks that problem without making you write code. It's a Windows desktop app that wraps the imgaug library in a point-and-click interface: pick from blur, color, geometric, weather, and corruption-style transforms, and it produces a batch of augmented images with matching filenames. The bundled documentation links each augmenter to its imgaug reference for deeper tuning. The project is Chinese-language, with video tutorials on Bilibili, so non-Chinese readers should expect the docs in Chinese even though the interface itself is visual. If you already script your preprocessing, the GUI will feel slower than a code pipeline, but for one-off dataset expansion or visual configuration it's a practical shortcut.