Seedance 2.0 - Advanced Workflow Series 8. Bypass Real Human Face Block Upload realistic character references to Seedance without triggering the face detection block. From character sheets to a close-up as the first frame. This video is an extension of the method shared by WTR Workflow + Prompts 👇

1. The secret: Make an infographic / collage with your realistic images - KEEP THE SIZE OF THE FACE SMALL COMPARED TO THE CANVAS The realistic image filter is based on detecting human faces. There is a pixel limit where Seedance's face detector doesn't trigger. In a 2K image, the maximum size I've tested that allows a face is around 200 pixels. However, in a 6K image, the maximum size increases up to 800 pixels. The key is to surround the face images with text and instructions so that Seedance reads the image as an infographic and not as a picture of a person. In this video, I first used an infographic explaining the prompt step by step. However, I tried making a collage putting the images with the name underneath as references, and it also worked. That is: we can use a single 6K image with small images inside with their respective names and reference them in Omni mode as if they were separately uploaded images. Shoutout @PocketScreenAI , that shared the original post of @wtry1102 and let me know how to approach the human face blocking

2. Create an infographic For this video, I created several references of myself. I generated an influencer-style image in MJ and then replaced the face with Nano Banana. Then I edited it to change the clothes with a hat, a Hawaiian shirt, and a beanie. Finally, I created two logo images. I used Gemini with its image tool to ask it to make an infographic-style image with instructions for Seedance, putting all the necessary images to create the infographic as input. Prompt: Create an infographic with the character's character sheet, stating that the first frame is 1.png, where the image 2.png appears with the title under image2 and where the forbidden face symbol has the subtitle SimbolBad and the image symbol with the green tick has the subtitle SimbolGood. In the infographic, an explanation will appear with the instructions for the video model (prompt) (ALL THE PROMPT IN ENGLISH; INCLUDING DIALOGUE). Where it indicates that it starts by first frame with a person speaking who says to camera "maybe you are tired of uploading an image to seedance and it does this" and right after the character points with a raised finger at the SimbolBad symbol floating in the air, (then makes the symbol disappear) then says "you will go to this" and then it appears on the same index finger pointing at the floating green check symbol. Then the character says: "Keep reading" and then transforms into the image where he is wearing a Hawaiian shirt (reference to the infographic's subtitle) "and create videos with real characters"... Then there is a cut and an image of him appears with a holographic VFX style sign and then he smiles at the camera in a close up.

3. Create a collage I used Photoshop to create a 5973x3360 px canvas (6K - 16:9). If you don't have Photoshop, you can create the same canvas in Paint, which is free and comes installed by default on Windows. In it, I introduced a simple text instruction "take this image as first frame" and then an arrow pointing to what I wanted to be the first frame. I tried to scale the first frame image to the maximum without it being rejected. The maximum size Seedance allowed for a close-up was 799 x 450 px. When I used 2K images with smaller sizes, Seedance had difficulty maintaining my face's consistency, but by scaling the canvas I managed to get it to read an 800px image, enough to preserve the details. I also tried generating the infographic video with a collage with references. It understood the references, although I didn't continue refining the prompt to get results since I liked the infographic one.

4. Introduce the collage/infographic into Seedance and Reference the image titles in the prompt You will put a name under each image in the collage. Then, you will reference each name in the collage indicating if you want the image to be a frame, or to be taken as a character sheet, or as a prop. In my case, I wanted to generate a close-up video with a first frame. This is the prompt: Character speaking to camera. Dialogue: "You did not believe me? I told you, (sighs) this image is the first frame. (pause) Take a look at the post, come on" Take the image pointed at in the collage as first frame, do not show marks or white background Upscale first frame to 4k before starting generation, preserve original character traits VOICE: LOW AND DEEP Prompt for the Beach Video: " " (Yeah, as info is in tan infogrpaphic there was no need for a text prompt) However, I tried another version with a Collage and references: Alternative Prompt: "FIRST FRAME" = FIRST FRAME of video 1. Video starts with FIRST FRAME a person speaking to camera in white plain background, dialogue: Perhaps you're tired of uploading an image to Seedance and it doind this..." 2. The person points to a symbol of a fece with his finger, "symbolbad" as frame raference 3. Dialogue "but with me, you will pass to this.." and symbol on the finger changes to the green tick of "symbolgood" 4. Transition (cut to morph) to character wearing hawaian shirt of "hawaian". New dialogue "Keep reading..." static holographic screen of "beach" appears "... and create videos with real character" 5. Zoom in to a close up of the face smiling

or just add a grid over it.... found out this works for images and videos, better for videos in my opinion i built a tool quickly if you want to try:

@Jaminel_Igwe Another gem from voxel 🔥

@Jaminel_Igwe Keep creating!!

Works well with digital art.

Is this a question or your opinion after testing it?

Thanks you so much!! I'm really thrilled that you've taken my workflow and expanded it so thoughtfully with such a detailed explanation 🙏 Making realistic characters so much easier to use in Seedance 2.0 is just awesome! The video and thread were incredibly clear and easy to follow — I learned a lot! This is going to let everyone create with realistic faces way more freely now. Let's keep making cool stuff together

Your method is awesome!! Thanks a lot for sharing it!!

Mastering the workflow. Seedance 2.0 has no limits if you know the path!

Working with resl references in Seedance is the end game

This is great ser. Would appreciate a video on workflows for organised advanced scenes with multiple shots and multiple actions, like conversation, Fight, dynamic actions, all happening in one environment. A format would be welcomed 🙏

tested the workflow, and unfortunately still got my requests blocked. would appreciate some tips if there are any

Try lower the size of the file but keeping resolution. Make faces smaller related to canvas. If not: add a little bit of blur or noise in photoshop to the faces

will try, thank you

OK, tell me if it works please

interesting workflow but feels risky sometimes

Thanks for taking the time to type down all that info. A lot of work was put into this! Respect.

I need to try that.