Universal replacement!!! Replace e-commerce products, clothes, shoes, faces, limbs, landmarks in scenes, etc. Keep 99% of the details unchanged, change perspective, and adjust lighting and shadow based on the background environment.

High-resolution image details, streamlined plugins, clarity.

General idea:

Mainly centered around the Flux Fill and Redux two redrawing models, merging product images and background images into unified elements. Ace assists in refining the generated images (to prevent detail loss such as text, lace details, etc.).

Consistent image proportions ensure better merging of product images. Sampling two merged images together allows the final generated image to better learn the details of the product image, preventing detail loss.

Steps:

1. Import the product/any image and the scene image. (The scene image requires masking the area, and try to ensure the masked area matches the product's spatial proportion in size.)

2. Fill in the prompt words for segmentanthing (for higher controllability, you can replace it with a node that doesn't require inputting prompt words).

3. Describe the expected generation prompt in short sentences. (For example: There is a product image of a bottle of perfume; the scene image is a beach environment. Simply describe: "A bottle of perfume placed on the beach.")

Everyone is welcome to use it, learn together, and provide valuable suggestions! The nodes are simple, without unnecessary complexity, making it suitable for reference.