Doubao (a type of Chinese dumpling) makeover video tutorial: 3 free steps to replicate using AI.

I've recently seen a bunch of transformation videos on Douyin, all of which have gone viral with at least 10,000 likes.

After careful consideration, I found that this format is very suitable for use.AICome to "Evil Cultivation Account" - only basic editing is required, and you don't need to show your face.

I have figured out a shortcut that can be replicated.

twoAIWith these tools, even beginners can easily replicate and create stunning video effects.

The faucet transformation video can be broken down into two parts:

The first part is when the character discovers the faucet and reaches out to turn off the water. In this part, we use the first and last frames to precisely control the character's movements.

The second part is a demonstration of the transformation; the video can be generated directly from the first frame.

We need to create three storyboards in this shot order.

For example, to make a video of a bean bun transforming into a character, our first step is to find a picture of a bean bun as a reference image, and then input the following...Prompt wordsThe storyboard was generated showing Doubao discovering the faucet indoors.

Prompt wordsThe girl in the picture is facing the camera. The background is a warm yellow wall. It's an indoor setting with ceiling light strips illuminating the room, but the source of the light is not visible. The camera is positioned above the sink, slightly tilted upwards. In the foreground, directly above the center of the image, a faucet is visible, with a stream of water flowing out, obscuring about 1/10 of the frame. The aspect ratio is 3:4.

The first image is now complete:

Using the image above as a reference, have Doubao perform the action of turning off the water.

Prompt wordsThe girl walks closer, her right arm extending out of the frame.

Using the image above as a reference, generate the transformed image.

Prompt wordsEliminate water flow, everything else remains the same.

Prompt wordsHigh-definition details and sophisticated color schemes are used to give the girl a fresh, girl group makeup look: dark gray contact lenses, long eyelashes, natural-looking brows, rose pink lip gloss, realistic hair texture, large wavy curls, and smooth, shiny hair. Her outfit is changed to an exquisite white strapless dress, complete with an Italian-style haute couture gemstone necklace. The dark purple background creates a mysterious and captivating atmosphere. The spatial layout and the girl's pose remain unchanged.

We use the three storyboard images generated above to create two videos in the Dreamscape.

Use the first and last frames to generate images; the reference images are the first and second generated images.

Prompt wordsThe girl looked at the faucet, took two steps forward, and raised her arm, extending it beyond the screen.

This gives us the first video:

The second video segment was generated using an image of the person after changing clothes as the first frame.

Prompt wordsThe atmosphere is captivating; the slow-motion shot shows the girl looking down slightly at the camera, exuding charm, with a strand of hair gently swaying in the air.

Second video:

Import the footage into CapCut and drag it into the video track. Add zoom and camera movement effects between the two video clips.

Search for the sound of water from a faucet in the sound effects library and add it below the first video clip. Search for the sound of a faucet changing clothes in the background music library, add the same background music, align the drum beat of the music with the changing clothes moment, and then export.

Ultimately, we got this transformation video:

Here's another interesting style to share: this version includes me in a costume video.

Here we need to first create a smoky-style image of a bean curd bun. The reference image is the second one generated from the bean curd bun above.

Prompt wordsThe girl in the picture wears smoky makeup, a cool expression, and a trendy, youthful style with colorful dreadlocks and a different outfit. The background is a mix of colors and dark tones, creating a specific atmosphere. The spatial layout and the girl's pose remain unchanged.

This will give you a smoky eye makeup look:

Then upload the generated image and my own avatar to Doubao to create a group photo storyboard.

Prompt wordsThe style of the girl in Figure 2 is changed to that of Figure 1. The girl in Figure 2 is added to Figure 1. The girl in Figure 1 is slightly turned to the side, and the girl in Figure 2 wearing VR glasses leans against the girl in Figure 1, touching her. The two of them are facing the camera and taking a photo together. The scene in Figure 1 remains unchanged.

This is how I got the photo of me and Doubao:

Finally, using the "first and last frames" function of the JiMeng app, I uploaded the two images generated by Doubao to create a video.

Prompt wordsIn Figure 1, the girl strikes a pose, and a girl in white outside the frame steps into the shot, waves to the camera, and the two pose for a photo together, the scene remaining the same.

The final video output is as follows:

Finally, use CapCut to edit them together, add sound, and you're all set!

This is very...SimpleHowever, my intention wasn't just to teach everyone how to make this video; I mainly wanted to share my experience with it.AISome ideas for implementation.

I think there are too many people nowAIThe implementation strategies are all wrong; everyone is thinking about doing something new, something that hasn't been done before.

But where are all those good ideas? The best way for ordinary people is to transform existing best-selling products into...AIProduced.

Creating proven best-selling products may not be the best approach, but it's certainly the most reliable one.