OpenArt
OpenArt
Workflows
Go to OpenArt main site
Upload workflow

Product Photo Relighting v2 - From Pre-existing photo, generate background, relight, keep details

5.0

0 reviews
52
7.8K
2.5K
4
Description

IMPORTANT

This is a workflow we created during this livestream:  https://www.youtube.com/watch?v=xjy3JyaPfHQ

It's a all in one workflow, where the user can:

- either generate a product, or start from a pre-existing product photo

- segment out the product through a SAM group

- generate a new background

- blend the original product on top of the generated background

- relight through mask (either a pre-existing light mask or by masking the resulting original product + background image in the Preview Bridge node)

- OPTIONAL: keep finer details (such as text) by a series of masks and post processing nodes

I personally would rather preserve details by using a frequency separation technique in Photoshop, as highlighted here:  https://www.youtube.com/live/xjy3JyaPfHQ?si=joT19pYsm30Zs2X9&t=2834

But if you want to do everything inside of comfyUI, I guess you can, it's just more tedious and time consuming.

Please watch the video to better understand how it all works.


Want to support me? You can buy me a coffee here: https://ko-fi.com/risunobushi

Cheers!

Andrea

Node Diagram
Discussion
M
Matt10 months ago

love it!

b
big dato (hairstyle)10 months ago

Traceback (most recent call last): File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) File "C:\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) File "C:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-post-processing-nodes\post_processing_nodes.py", line 556, in dodge_and_burn dodged_image = self.dodge(image, mask, intensity, mode) File "C:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-post-processing-nodes\post_processing_nodes.py", line 574, in dodge return img / (1 - mask * intensity + 1e-7) RuntimeError: The size of tensor a (1024) must match the size of tensor b (64) at non-singleton dimension 2

a
andrea baioni10 months ago

What's this error in response to?

Are you trying to relight from a custom mask in a preview bridge node or from a custom mask imported in a Load Image node?

Also, please refer to the complete rework v3 here:  https://openart.ai/workflows/risunobushi/product-photography-relight-v3---with-internal-frequency-separation-for-keeping-details/YrTJ0JTwCX2S0btjFeEN

Author

a

andrea baioni

33
123.2K
1.9K
365.4K

Reviews

No reviews yet

Versions (2)

  • - latest (10 months ago)

  • - v20240517-122408

Primitive Nodes (19)

Note (18)

Reroute (1)

Custom Nodes (56)

ComfyUI

  • - CLIPTextEncode (6)

  • - CheckpointLoaderSimple (2)

  • - VAEDecode (3)

  • - PreviewImage (9)

  • - VAEEncode (2)

  • - MaskToImage (4)

  • - ControlNetLoader (1)

  • - KSampler (3)

  • - EmptyLatentImage (2)

  • - ControlNetApply (1)

  • - LoadImageMask (1)

  • - ImageBlur (1)

  • - LoadImage (2)

ComfyUI Essentials

  • - MaskFromRGBCMYBW+ (1)

  • - ImageResize+ (1)

ComfyUI Impact Pack

  • - PreviewBridge (5)

  • - SAMLoader (1)

  • - ImpactGaussianBlurMask (1)

comfyui-art-venture

  • - ColorCorrect (1)

ComfyUI-IC-Light

  • - ICLightConditioning (1)

  • - LoadAndApplyICLightUnet (1)

ComfyUI-post-processing-nodes

  • - DodgeAndBurn (1)

KJNodes for ComfyUI

  • - GrowMaskWithBlur (1)

  • - RemapMaskRange (1)

segment anything

  • - GroundingDinoSAMSegment (segment anything) (1)

  • - GroundingDinoModelLoader (segment anything) (1)

WAS Node Suite

  • - Image Blend by Mask (2)

Checkpoints (2)

epicrealism_naturalSinRC1VAE.safetensors

LoRAs (0)