👗If you want to make stylish showcase videos, this is it! BandyAI, where 'what' and 'how' you want to showcase your products is already 'laid out' | A single image makes the model move and showcase automatically! Includes a guide on prompts for effective showcasing | AI Video Generation

Welcome, thank you for coming🐲
I'm Natsume, an AI image generation and AI video generation researcher and creative writer🐲

This time, it is theAI agent specialized for e-commerce sites

Bandy AI

introduction, hands-on review, and the release of created works, which is thethird installmentof this series.
Well, everyone, what would you think ifjust one 'model image' could move 'automatically' to beautifully showcase a product?What do you think?

Huh...? ( ^ω^ )?

Huh...? ( ^ω^ ) It moves this time!?

It moves!!✨
The AI that can perform such magic is this one!Bandy AI!

※This article is a PR article created at the request of Bandy AI (*´ω｀)

Last time, as 'image generation', I introduced it with the content thatyou can easily generate high-quality, high-resolution 'model x product x background' images just by choosing a model image for the product,.

Well, this time

that image generationmoves!!

👗Video generation where the model moves

And this time, I will introduce one more thing in image generation
👠Pose generation & angle change image generation

This is a two-part introductory article (*´ω｀)✨

So, how exactly do you operate it, and what is it like?👀✨
I hope you will take a look with me, Natsume (*´ω｀)✨

👓Bandy AI's 'Tools'

Bandy AIOn the homepage, use the Tools section on the left side of the screen.

In the previous article, I introduced the three features in the red boxes.
This time, I will...

👠Generate Poses & Change Angles
👗AI Product Video

I will structure the article in this order!

Well then, let's get started!✨

👠Generate Poses & Change Angles Image Generation

Generate poses, change angles.
This means it's an image generation tool that can 'change' the 'pose' of the original image.

I see, I see(*´ω｀)
A photo of a model wearing the product in advance, oh, it seems a 'mannequin' is also fine(*´ω｀) This is also an amazing point✨

Let's press 'Select Pose'.

Oh~(*´ω｀)
First, let's look at Select from presets'.
Well, I see, is this what you call a 'posing template'?

And, look here, please pay attention to the red box.
Right now, I haven't touched anything.
It says 'plus' and '4 images' here.

And if you...
'increase' this...

Do you see what I mean?

・1 full-body image
・1 upper-body image

It means it will generate these for you👀✨

Let's try generating with this image model.

Oh, the original image was mostly upper body, but since I wanted it to focus more on the product, it gave me a bit of a close-up.
For the full body, since what I want to 'showcase' is the clothing and not the shoes, I think this is a very appropriate composition✨

Next,
I selected 'Enter manually,' which is next to 'Select from presets.'

There were a few ideas in the suggestions, so I chose one.

Seated pose with hands on hips
(Seated pose with hands on hips)

I generated it with this.

Ah... *sigh* ( ^ω^ )

Looks like you weren't satisfied with that... (*´ω｀)

Natsume-shi ran away to another model, and after thinking about the pose...

(´;ω;｀) Waaah💦

What is a pose??
How can I make it look better?
Is this what Bandy AI does normally? Isn't that amazing!? (´;ω;｀)

Nunununu (*´ω｀)
I can do it too!!!🔥

I also tried to make the hair dance a bit.
Regarding this, Natsume thought 'Hmm?' at the phrase 'enter it yourself,' and it might just be a coincidence, but

It doesn't actually say 'pose' in the phrase 'enter the 'pose' yourself.' It's not restricted to just poses.

This means I thought I could mention the model's expressions and atmosphere.

If you look at the screen one more time,

The text field for 'enter it yourself' actually holds '1500 characters.'
So, this is what corresponds to the 'prompt' that makes the model operate (*´ω｀)✨

And so, that is how you get an image like the one above, where she elegantly showcases the dress with a smile (｀･ω･´)ゞ✨

Please feel free to use this prompt exactly as it is, or modify it as you like; it is yours to use freely✨

Now, everyone, go ahead and 'show it off!!'✨(๑•̀ㅂ•́)و✧

👗AI Product Video

Now, let's move on to the main part of this article: 'Video Generation'.

I see, I see. After using it, I've gotten a general idea of the UI design.

Okay, I'll go with the girl with the short bob(*´ω｀)
Next up is 'Presets'.

I see, it says 'Please select a video style'.

When you click 'Please select a video style', as shown above,

a list of videos will appear.
*The screen is a screenshot, so it is static.

So, I will choose this woman who is showcasing it so sexily.

And, I see, the next step is choosing the settings needed to create the 'video'.

Here, Natsume chose:
✅ Standard
✅ 10 seconds
✅ 1 video

I have selected these.
Now, let's generate it! 👀✨

This here is the movement selected from the 'Preset' video movements.

She was really moving like this.
And it is,the model from the image moving properly.

Wow, such smooth movement. Wait, was creating videos always this easy? Huh?

Next, let's look at the 'Explain' button next to 'Preset'.

This was included in the defaults from the beginning.

1500 characters... this is a 'prompt', isn't it? (*´ω｀)💡 Ping!

According to my research, the timing for when that 'description' above 'enters' iswhen you select the image under 'Try out these examples'.

In other words, each of these five 'examples' has

This means that the 'prompts' for the video are already prepared.

That is how it works.

I tried changing it to a different image.

As expected, the content of the description was different.

Exactly, since these are videos meant to 'showcase' each piece of 'clothing,' it wouldn't make sense if they weren't different, right? (*´ω｀)

When you use it, you pull images from your own folder, right?
When you upload them, there is nothing in the 'describe' section.
The only things in there are the 5 'example' images.

So...💡

I think it's best to check which of these 5 examples is closest to your own product to see how it handles things,copy the description for that similar example, paste it into a memo, and then import your image.

And then, why not try modifying that description, or if the clothing is the same shape, try using it as is? (*´ω｀)✨

I want to show you at the end that this default 'description' is at a level where it doesn't even need modification, so please take a look.

Well then, for the lady in the shorts,I generated a video using the description that was in the 'default'.

I see, I see!

Ah... personally,

I also want to see the back, the back of the shorts.

So, I haveaddeda little bit to the 'describe' prompt.

This one, it's a failure where she ended up facing forward before I realized it.
It's the AI's cute magic trick of deception (*´ω｀)✨
Although it's a failure, this is exactly what the AI was instructed to do.
It showed various angles, and at the end, it went 'whoosh!' but she is facing forward, after all.

Hmm... (*´ω｀)
Was the 'front → back → front' instruction good?
Should I have specified how much 'rotation'?

I wonder if human verbalization is surprisingly 'vague'?
Lately, video generation has been recognizing that.

Videos require clear 'instructions' for the beginning, development, turn, and conclusion, but I think humans are capable of expressing that loosely with just a vibe!! ✨(*´ω｀)

Now, now, 'improvement'! 'Improvement'!!

This is where the 'human' (me) comes in! (*´Д｀) Ta-da!

I'll change the instructions, which means I'll change the 'wording'.
First, let's see what happened when I changed the 'wording' 👀✨

It was a success (*´ω｀)

To the AI:
・First, rotate 360 degrees
・Do not cut, do not jump, rotate with the whole body
・Show the 'front', 'side', and 'back' naturally during the rotation

I tried being specific (*´ω｀)
I forgot to say 'face the front' at the end, but maybe the emphasis on '360 degrees' worked, as it ended facing the front ✨

The previous prompt might have been too rough.
In the middle of that, the camera focused on the pants quite well.

This is what an 'AI product video' becomes ✨

By the way...
At first, I didn't know the specifications, so I did both 'preset' and 'describe', meaning,I selected a preset video + entered a prompt to generate it(*´ω｀)

Then, 'describe' won.
The way of saying it was, well,'Describe' was prioritized.

Preset < Describe

BandyAI is

・Default image or video (fixed)
・What to do with the image or video? Input-based description (free)

Since it's a choice between two, even if you accidentally do both,

the input type (free) takes priority.

That's how it works (*´ω｀)✨
However, it might cause bugs, so if you are reading this article, I would appreciate it if you could keep in mind to select one or the other (*´Д｀)✨

👚Finally

How was it?

Well, I'm the type who just 'wants to touch the prompt' and 'wants to go on an adventure by changing the prompt,' so I have been tinkering with it, but honestly,

for AI product videos, 'modifying' the 'explanation' prompt is unnecessary!!

Huh? The reason... well...

Now, here is a question! (Sudden!?)
Out of the next two videos, which one is the video with the 'prompt created by Natsume'!!

[1]

[2]

OMG!! (*´Д｀)
Please don't say that the answer is already obvious from the YouTube title!

for AI product videos, 'modifying' the 'explanation' prompt is unnecessary!!

The reason for this, which I realized when I tried making one myself, is that BandyAI is

creating videos with a focus on the 'outfit,' or in other words, the 'product'

(*´ω｀)

With this in mind, I want you to watch it one more time,

Natsume's video.
This is already becoming a 'standard' for Pollo-san

Close-up punchline

Well, you see, I'm sorry, it's just my obsession (*´ω｀) Something wrong?
This 'habit' is just stuck with me💦

In other words, my videos

focus on the 'model's' facedon't they?

In contrast, BandyAI videos

slowly close in while focusing, focusing on the dress

as they do it.

I realized that since 'what' you want to show

is already 'placed' by default, there is no need to correct it at all.

Honestly, BandyAI has so much to 'learn' from.

I, Natsume, have been exploring various AIs and continuing to study their 'habits' and how to deal with them.

But really, this is the first time I've encountered an AI where I can 'understand' and 'feel' the 'creator' (person) on the other side through the AI.

I can feel the various considerations for the user.

'Professional quality for anyone'
'Easy to do'

They truly walk the talk; all you have to do is select an image and a model.

That is the kind of 'world' we are in now.

To those of you running your own businesses, or those thinking of starting one.

Why not try using this user-friendly BandyAI for your e-commerce site strategy? (*´ω｀)✨

That was the AI designed to 'captivate' people, Bandy AI❣

Thank you so much for your likes, comments, adding to magazines, and using Min-Photo🥰
Please continue to support me in the future.

Wishing you a wonderful AI video life~💐
*This article is co-written with Bandy AI.
*The article's header (catch) image is not related to Bandy AI.

🔽Bandy AI Magazine

#AIVideo
#AIVideoGeneration
#VideoGenerationAI
#GenerativeAI
#BandyAI
#AIBusiness
#AISideHustle
#AIMarketing