🌟 Model Introduction
This is a highly practical LoRA specially designed for e-commerce, apparel display, and AI-assisted photo retouching.
It can accurately identify complex real-life photos containing people, remove the human model and environmental background with a single click, retain only the pure clothing itself, and generate clean product images with a pure white background. Whether it is everyday clothing or complex Hanfu (traditional Chinese attire), it perfectly erases any trace of the "person," giving the clothing a natural drape as if it were laid flat or hung in a photography studio.
✨ Core Features & Supported Types
This model has been specially trained and currently perfectly supports the precise extraction and generation of the following 4 types of items:
👚 Tops
👖 Bottoms
👗 Full-body Outfits/Dresses
👜 Bags
🛠️ Parameter Settings & Usage Tips
For optimal results, please refer to the following parameter recommendations:
Trigger Words: `生成图中人物(全身衣物|上半身衣物|下半身衣物|包包)的商品白底图，去除背景，去除人物，只保留衣物` (English meaning: Generate a white-background product image of the person's [full-body outfit | upper-body clothing | lower-body clothing | bag] in the picture, removing the background and the person, keeping only the item)
LoRA Weight Adjustment — Core Tip ⚠️:
Weight = 1.0: The model will exert its maximum effect. At this level, it will not only remove the person but also smooth out the natural wrinkles on the clothing, presenting an extremely flat, "absolutely commercialized" effect as if it had been perfectly ironed (suitable for graphic design or pure display).
Weight = 0.6 ~ 0.8: If you want the clothing to retain the natural wrinkles, realistic lighting, and material textures left from being worn, it is strongly recommended to lower the weight appropriately. This will make the product image look much more authentic and natural.
💡 Use Cases
E-commerce sellers quickly creating main images for product detail pages.
Fashion designers extracting inspiration and documenting clothing patterns.
AI art enthusiasts preparing materials for clothing-swap or virtual try-on workflows.
Trigger words
You should use 生成图中人物全身衣物的商品白底图，去除背景，去除人物，只保留衣物 to trigger the image generation.
