Генерация Картинок Нейросетью. Bing Image Creator Как Пользоваться?

Summary

This video tutorial introduces Bing Image Creator, a Microsoft service that generates free images based on user prompts. Viewers learn how to access the service, change interface languages, and navigate through idea overviews and personal creations. The video demonstrates the process of generating images, including the use of 'boosts' to speed up the process. It also covers the limitations, such as the inability to generate images with celebrities or control the aspect ratio and number of outputs. The speaker shares tips for refining prompts to achieve better results and discusses the service's pros and cons, highlighting its ease of use, language support, and free access, while noting its restrictions on commercial use and photorealism.

Takeaways

• 🌐 Bing Image Creator is a free service from Microsoft that generates pictures based on user prompts.
• 🔑 Users need to log in with their Microsoft account to start using the service.
• 🌍 The service supports over 100 languages, allowing users to change the interface language through the settings.
• 🖼️ Users can view and download pictures created by others, but should note that images are stored for only 90 days.
• 💼 The 'Creations' tab displays pictures that the user has recently generated.
• 🚀 Users are given a certain number of 'boosts' or 'accelerations' each day that speed up the image generation process.
• 📝 To generate images, users must describe the object and style they want, with the service supporting a range of styles, though not an exhaustive list.
• 📐 The neural network generates images in either a square or 3:4 ratio, and the number of images produced can vary from one to four.
• 🎨 The service can write text on images, but the quality and accuracy of the text may require precise and clear prompts from the user.
• 🚫 There are limitations, such as insufficient photorealism, no commercial use of generated images, and no generation of images with celebrities.
• 🆓 Despite the limitations, the service offers the advantage of being free and easy to use, with the potential for generating good results with enough effort.

Q & A

What is Bing Image Creator and how can one access it?

Bing Image Creator is a service from Microsoft that allows users to generate pictures for free. To access it, one needs to go to the provided site, log in with a Microsoft account, and join the service.

How many languages does Bing Image Creator support and how can a user change the interface language?

Bing Image Creator supports more than 100 languages. Users can change the interface language by going to the top right menu, selecting the settings, and then choosing the Language tab to select a more convenient language.

What are the two main tabs found on the Bing Image Creator interface?

The two main tabs on the Bing Image Creator interface are 'idea overview' and 'creation'.

How can users view and interact with pictures created by others on Bing Image Creator?

Users can view pictures created by others on the 'idea overview' page. They can click on a picture to see the industrial design it had, remake or supplement it, and generate another picture based on it. They can also save or download the picture to their computer or collection.

What is the duration for which pictures are stored on Bing Image Creator?

Pictures generated on Bing Image Creator are stored for a maximum of 90 days, after which older pictures are deleted.

What are 'boosts' or 'accelerations' in Bing Image Creator and how do they function?

Boosts or accelerations are tokens given to users on Bing Image Creator. Each picture generation requires one boost. These boosts help in faster generation of pictures, but if they run out, pictures will still be generated, albeit at a slower pace.

What kind of guidance does Bing Image Creator provide for creating industrial goods?

Bing Image Creator provides answers to frequently asked questions and tips on how to create industrial goods on the main page. It suggests describing the object, adding adjectives, detailing the image, and specifying the style at the end of the request.

What are the limitations regarding the aspect ratio and number of pictures generated by Bing Image Creator?

Bing Image Creator generates pictures either in a square format or a 3:4 ratio, and the number of pictures generated (from one to four) cannot be influenced by the user.

What is the process for creating a custom picture on Bing Image Creator?

To create a custom picture, a user needs to write a custom message in the designated area or click on the 'Surprise me' button, which allows the system to generate a custom message for the user to create a picture based on that prompt.

Can Bing Image Creator generate pictures with text on them?

Yes, Bing Image Creator can generate pictures with text. Users need to specify the text they want on the picture in their request.

What are some of the pros and cons of using Bing Image Creator as mentioned in the script?

Pros include the free nature of the service, the ability to generate images indefinitely, understanding of more than 100 languages, and ease of use. Cons include insufficient photorealism, no commercial use of pictures, fixed aspect ratios, and the inability to generate pictures with celebrities.

Outlines

00:00

🖼️ Bing Image Creator Overview

This paragraph introduces Bing Image Creator, a free neural network service by Microsoft for generating pictures. Users need to log in with their Microsoft account and can change the interface language from over 100 options. The service includes features like idea overview and creation, where users can view and recreate images generated by others. There's also a collections feature, but images are stored for only 90 days. Users receive daily 'boosts' to speed up image generation, and the interface provides tips for creating industrial goods and crafting precise queries. The neural network supports various styles and generates images in square or 3:4 ratio, with an unpredictable number of images per generation. The paragraph also touches on the 'Surprise me' feature and the process of creating a custom picture.

05:04

🔍 Experimenting with Prompts and Text on Images

The speaker discusses the process of refining prompts to achieve better image results with Bing Image Creator. They share examples of initial vague prompts leading to unsatisfactory images, such as 'hand and nail care' resulting in creepy hands. By providing more specific prompts, like 'Alice in Wonderland saw the Cheshire cat sitting on a tree,' the generated images improved in quality and relevance. The paragraph also covers the neural network's ability to write text on images, with examples of varying success. The speaker emphasizes the need for patience and precise requests to achieve legible text. They also discuss the neural network's limitations, such as insufficient photorealism and restrictions on commercial use, as well as its inability to generate images with celebrities.

10:08

📚 Summary of Bing Image Creator's Features and Ease of Use

In the final paragraph, the speaker summarizes the advantages and disadvantages of Bing Image Creator. They highlight the service's free nature, language support for over 100 languages, and the ease of use. The speaker also shares their experience with generating images for 'Gray Wolf and Little Red Riding Hood,' noting the adequacy of the results compared to other services. The paragraph concludes with a reminder of the service's limitations, such as the fixed aspect ratios and the inability to generate images with celebrities, but overall, it emphasizes the positive aspects and the potential for satisfactory results with careful prompting.

Keywords

💡Neural Network

A neural network is a set of algorithms modeled loosely after the human brain that are designed to recognize patterns. In the context of the video, the neural network is used to generate images by Bing Image Creator, a service provided by Microsoft. The script mentions how the service uses neural networks to create pictures based on user prompts, highlighting its ability to understand and generate complex images.

💡Bing Image Creator

Bing Image Creator is a service from Microsoft that utilizes neural networks to generate images for free. The video script describes how to access and use this service, emphasizing its user-friendly interface and the ability to create images in various styles and languages. It is central to the video's theme of exploring a tool for image generation.

💡Microsoft Account

A Microsoft account is required to access various Microsoft services, including Bing Image Creator. The script explains that users need to log in with their Microsoft account to start using the service, which is a prerequisite for generating images and interacting with the platform.

💡Interface Language

The interface language refers to the language in which the service's user interface is displayed. The video mentions that Bing Image Creator supports over 100 languages, allowing users to change the interface language to their preference, which enhances accessibility and user experience.

💡Collections

Collections in the context of Bing Image Creator are a feature that allows users to save images they have generated or liked. The script notes that images are stored for only 90 days, emphasizing the importance of downloading images to a personal computer to avoid losing them.

💡Boosts or Acceleration

Boosts or acceleration are tokens given to users daily that allow for faster image generation. The script explains that one boost is deducted with each image generation, but even if users run out of boosts, the service will continue to generate images, albeit at a slower pace.

💡Industrial Goods

Industrial goods refer to the types of products or items that are manufactured for use in the production of other goods or services. The video script provides an example of creating an industrial product image, such as a bouquet of roses, using Bing Image Creator, showcasing the service's capability to generate images of various subjects.

💡Photorealism

Photorealism is a style of art where images are created to closely resemble photographs. The script mentions photorealism as a style option in Bing Image Creator and discusses the quality of the generated images, noting that while some are impressive, others may lack the desired level of realism.

💡Prompts

Prompts are the textual descriptions or requests provided by users to guide the neural network in generating images. The video script discusses the importance of crafting clear and detailed prompts to achieve better results, using examples such as 'care for hands and nails' to illustrate how precise prompts can lead to more accurate image generation.

💡Commercial Use

Commercial use refers to the utilization of a product or service for monetary gain or business purposes. The script points out that Bing Image Creator does not allow for the commercial use of generated images, which is a limitation for those looking to sell images or use them for commercial products.

💡Aspect Ratio

Aspect ratio is the proportional relationship between the width and height of an image or screen, typically expressed by two numbers separated by a colon. The video mentions that Bing Image Creator generates images in either a square format or a 3:4 ratio, and users cannot influence this aspect ratio, which is a limitation of the service.

Highlights

Introduction of Bing Image Creator, a free neural network service by Microsoft for generating pictures.

Instructions on how to access and log in to Bing Image Creator using a Microsoft account.

Service supports over 100 languages and allows users to change the interface language.

Overview of the user interface with tabs for ideas and creations.

Ability to view, save, and download pictures created by other users.

Pictures saved on the service are stored for only 90 days.

Daily allocation of 'boosts' or 'accelerations' for faster picture generation.

Frequently asked questions and tips on creating industrial goods are provided.

Guidance on creating precise queries for generating images.

List of supported styles for image generation, though not exhaustive.

Images are generated in either square or 3:4 ratio, with no control over the number of images produced.

Demonstration of creating a custom image using a prompt.

Example of generating an industrial product image with photorealism.

Discussion on the trial and error process of refining prompts for better image results.

The neural network's capability to write text on images.

Analysis of the pros and cons of Bing Image Creator, including free usage and language support.

Limitations such as insufficient photorealism and restrictions on commercial use of generated images.

Inability to generate images with celebrities due to copyright issues.

Conclusion emphasizing the ease of use and potential for achieving good results with effort.

Transcripts

Hello everyone who is watching. In this video we will look at another neural network that allows us

to generate pictures for free. This is a service from Microsoft called Bing Image Creator. in order to

start using it, just go to this site, the link will be in the description,

and here you need to log in with your Microsoft account. To do this, we click here Join

and Create, log in with a personal account. If you don't have a Microsoft account, you

can easily create one using email. I have a Microsoft account, so I just log in

and we get to this page. Here, first of all, we can change the interface language. This service

supports more than 100 languages. Go to the top right menu in the settings - Language tab,

and here select the language that is more convenient for you. Here, the interface has switched. at the top left we see

two tabs: idea overview and creation. a review of ideas is this page itself,

where we see pictures created by other users. We can go in and see

what kind of industrial design this picture had, we can somehow remake it, supplement it and generate

another picture based on it. We can save this picture - download it to our computer,

we can save it to our collection. If we click on collections, here is a separate tab,

here are the pictures that I once saved. You must remember that on this service pictures

are stored only for 90 days and old pictures are deleted. As you can see, they are no longer available,

the ones that I previously saved. Therefore, if you have generated some successful picture and

want to save it, then download it directly to your computer. in the creations tab, here are the pictures

that I generated recently. Also, during the day we are given a certain number of such tokens,

they are called boosts or acceleration. With each generation, one

boost is deducted from us. Thanks to these boosts, pictures are generated faster, but if they run out, it’s okay,

pictures will still be generated, just slower. here on the main page, if

we scroll down, there are answers to frequently asked questions, and here there are tips on how to create industrial goods.

How to create more precise queries. Here, of course, the instructions are quite short. Well, in general

terms, we need to describe the object - what exactly we want to see in the picture, add adjectives,

describe what the image should be, and at the end it is advisable to add what style it will be in.

As for styles, I asked a question in Microsoft chat, which is called Copilot and works

on the basis of gpt chat. What styles does Bing Image Creator support? he gave me a list,

although this list is not complete - there are only eight styles. These are just a few of the ones

it supports. and of course, no one forbids us to experiment and

add to our request those styles that we know. Another important note - here

this neural network always generates pictures either square or in a 3:4 ratio,

as a rule there will be four pictures as a result of generation, but sometimes it produces three pictures,

and can even produce only one picture. We cannot influence this process in any way. Therefore, when

generating images, it makes no sense to set the aspect ratio; it will still generate either a square

or a 3:4 ratio. and in order to create our own picture, we need to write a custom message here, or

click on the “Surprise me” button - the system itself will create some kind of custom message. and let's see what

is offered to us here. create a picture based on this prompt. generation is underway. Here's a funny dog

​​in Digital Art style. let me introduce some industrial product. I have prepared an industrial project “a bouquet of bright red roses

in a vase, photorealism” and let’s try to create it. Well, these are some pretty nice pictures. enlarged it

is very beautiful. and I’ll say a couple more words about promtah. using the example of those

industrial items that I have already created. I had this short request for “hand and nail care.”

As you can see, he understood the meaning, but the pictures turned out ugly, there is too much here,

if you look closely at the details, then it’s actually some kind of nonsense. Then I clarified

this promt, wrote “care for hands and nails, only two hands, photorealism.” He created three

such pictures for me. Two of them, of course, are not suitable, if you look more closely, here are some

problems with the fingers, it generally looks creepy here - it looks like hands in gloves,

but pay attention, it’s not clear where the skin ends and the gloves begin, it’s

just generally creepy. So, not everything is so perfect here with this creation of pictures. Well, one

picture turned out to be very decent. Although if you're being picky, it's also not clear what kind of

instrument this is. Well, God is already with him. Another fabulous example. For example, here “Alice in

Wonderland” there are such pictures with a lot of details and, frankly speaking, they are also rather scary.

then I clarified “Alice in Wonderland saw the Cheshire cat sitting on a tree.” And here,

too, some of the pictures turned out well. Well, here, for example, look, Alice with a tail,

here some kind of horror has already come out, there’s something else, some kind of cat is peeking out. here it’s

more or less normal, but there’s something wrong with her hand. here more or less, but the picture itself

is not of very good quality. This picture is also more normal, but if you look closely, there are

so many small cats hiding in the bushes. Then I changed this promt and added “Alice in

Wonderland saw a smiling Cheshire cat sitting on a tree.” and here things

got more interesting. Here too, if I’m being picky, it’s like some kind of strange flying object,

just in case he made Alice smile too. This is the picture I liked the most;

there are the fewest artifacts here. Here the cat seems to have turned out well, but look what’s

going on around here - how many small cats are on the branches, and here there are two more Alices dancing. And here, too,

everything seems to be fine, some details are from this fairy tale, but there is some weirdo there too,

looking like Alice in a hat that looks like a mushroom with ears. Well, that is, what I’m saying is that

be prepared that not all pictures turn out good, and you need to experiment

with prompts - change words, substitute additional words, swap words. In

general, experiment with industrial products to achieve a good result. This neural network can also

write text on pictures. So I made the following request: “a jar of cream is on the table,

on the jar is the text: gentle care for the skin of your hands.” He created four pictures - two of them with text,

two without text. Let's take a closer look. He writes the text only in English. the first

phrase is all right - gentle care, but then it’s very illegible, and in my opinion some

nonsense is written there. second picture - everything here is legible, but with errors. Well, you probably need

to be patient, formulate your request more precisely, and then he will write normally. and two other

pictures turned out without text, but in fairly good quality. I especially liked

this picture . and let’s also talk about the pros and cons of this neural network.

I would consider the disadvantages to be insufficient photorealism. Look, this picture is a portrait of a girl,

to be honest, even in comparison with Leonardo, the photorealism here is insufficient, in my opinion,

it is still clear that this picture is drawn. Another significant drawback can be noted

that this service does not allow the commercial use of pictures. Well, for those

for whom it is important. for example, there are types of income such as selling pictures on stocks, or

printing on T-shirts. in this case, this service will not work. Another drawback, perhaps not

so significant, is that either a square picture or an aspect ratio of 3x4 is generated,

and the service will also generate for us from one to four pictures - we cannot influence either the number or

the aspect ratio. And another drawback is that this service does not generate pictures with

celebrities. Let's check. I will enter the query “Angelina Jolie” You see, he warns that

you cannot generate with celebrities. and if we take not the first and last name, but the role that

she played in the movie and try to create, well, you see, it’s no longer the actress herself, but her role.

one might say, vaguely similar to her. As for the advantages, this is, of course,

the freeness of this service; in fact, we can generate images indefinitely; even

when we run out of these boosts, it will still generate images, just more slowly. and

yet, with enough effort, you get pretty good results. Also

the service understands more than 100 languages. I also want to point out that I'm pretty understanding regarding

requests. So I generated about the Gray Wolf and Little Red Riding Hood - he understands the request well,

the pictures are quite adequate. If you watched my previous video, in the Leonardo service I tried

to create the same story with Little Red Riding Hood - I got some pretty strange pictures there. Well,

the service is also very easy to use, without any problems. Perhaps this is all I wanted

to say in this video, if it was useful, like, write comments and see you again.