Can you believe it?
As "0" AI ArtThe basic whites can also produce these amazing things from 0-80 pointsplaybill.
I'M SURE YOU'VE SEEN A LOT OF AI DRAWINGS ON THE MARKET, BUT THE RESULT IS NOT IDEAL。
AND TODAY WE'RE GOING TO BRING YOU A WHITE, VERY FRIENDLY, BABY-SITTER-LEVEL AI DRAWING COURSE。
THERE'S ONLY ONE TARGET! IT'S THE LITTLE WHITES THAT MAKE THE ZERO-AI PAINTING BASE, AND THEY CAN DO IT RIGHT AWAY, AND THEY CAN MAKE A WORK THAT CAN BE DIRECTLY "CROSSED."。
I. Tool preparation
Here we go
THE TOOL WE NEED TO USE IN THIS PERIOD IS THE LIGHT OF THE COUNTRY 'S MAP - THE DREAM AI。
Website:
- https://jimeng.jianying.com
When you come in, everybody log in。
Very well, free users can also receive 60-100 points per day at random。
If you're looking for something bigger, you can upgrade your membership。
II. Precision chart 2 techniques
One of the greatest advantages and one of the greatest disadvantages of the map is that it is too random。
What if we want to produce a certain kind of picture or poster precisely
AND I'M GOING TO TEACH YOU TWO WAYS TO TURN AI FROM AN ARTIST WHO LIKES TO PLAY FREE TO A PROFESSIONAL DESIGNER WHO CAN UNDERSTAND OUR INSTRUCTIONS WITH PRECISION。
First move: To the map
What do you mean
It's like we're on the Internet and we're looking at some amazing pictures or posters。
This one, for example, feels like I want it。
BUT HOW DOES THIS WORK WITH AI
I'll teach you a trick, you don't even need to know any drawing techniques, you just need to find a reference map and get AI to help us reverse Prompt。
BUT THERE'S A KEY POINT IF YOU JUST ASK AI:
"What's the hint of this picture?"
That might not work。
I'll share with you the all-powerful formula I've been using
# ROLE:
– Please act as a professional image analyst and hint engineer。
# TASK:
– Your task is to analyze this reference picture I uploaded
AND TO GENERATE ME A DETAILED, STRUCTURED PICTURE, WITH THE AIM OF MAXIMIZING THE AI DRAWING TOOL
Reverts the style, content and structure of the original map。
# ANALYTICAL DIMENSION
, please output the results according to the following structure and requirements:
1. Core subjects:
It's the most important person, object or focus in the picture
They include details of their appearance, gestures, faces, clothing, etc。
(Soft space limited, small partners in need request personal contact)
For example, we want to create a picture of this feeling。
AS YOU CAN SEE, THE FOLLOWING FIGURE IS THIS FORMULA THAT GUIDES AI FROM Subject, style, structure, light, colour and detail Eight dimensions, go carefully and analyze the picture。
LOOK AT THE RESULT GENERATED, THE LEFT IS A REFERENCE DIAGRAM, AND THE RIGHT IS GENERATED BY THE REVERSE HINT AI。
ALTHOUGH IT'S IMPOSSIBLE TO DO THE EXACT SAME 100% BECAUSE OF THE RANDOMNESS OF THE AI, BUT THE IMAGEStyle, structure, atmosphere and core elementsIt's a perfect recapitulation of our reference
IF YOU DON'T THINK IT'S CLOSE ENOUGH, THEN YOU'LL BE ABLE TO PUT THE IMAGES AND REFERENCE MAPS TOGETHER SO THAT AI CAN FIND THE DIFFERENCE。
I'm very kind enough to get you all ready。
After the test, I put you out of this hint
IT'S NOT VERY EFFECTIVE TO APPLY IT TO THE AI PRODUCTION TOOL
TO ENABLE 100% TO REPEAT REFERENCE DIAGRAMS。
I'm going to upload a picture of you from the reverse tip
Please make a detailed comparison between this and the reference
Find a gap between the two maps and point out specific problems
one by one, output in markdown code box format
Final output optimized hint。
AI FINALLY RE-EXPORTS OPTIMIZED HINTS, COPYING THEM AGAIN TO DREAM CREATION。
In addition, I modified some of the objects in the picture with a local redrawing tool。
For example, if you change the word to "burn", if you change the washing machine to a stove on the right, if you change the pool to a flame, see if it works well
THIS IS THE FIRST STEP IN OUR PRECISE CONTROL OF AI'S FIGURE。
It's a very good way to do it when we have a clear imitation。
If you want to produce a picture of a better quality of text, I'm here to suggest a three-step approach。
The gentleman makes a picture, generates text information, then goes to the third-party photo editing software, and sets it up。
Second move: Structured biographic hint formula
But if there's only one idea in our minds, but there's no specific reference, and we want to control the image。
So what do we do
This is when we need to use the second approach, the formula that I sum up。
THE ESSENCE OF THIS FORMULA IS TO TELL AI IN A STRUCTURED WAY WHAT'S IN OUR MINDS。
The clearer and more structured you give it, the more it can read your mind and produce the results you want。
You can remember this all-powerful formula:
Core subjects
2. Scenario setting
3. Conceptual perspective
4. Light atmosphere
5. Artistic style
Colour tone
7. Text and graphics
8. Details and quality
Picture scale
IF YOU DON'T FILL IN THE FORMULA OR YOU DON'T FEEL GOOD WHEN YOU FILL IT IN, YOU CAN USE THE A.I. TOOL TO HELP US GENERATE THE FULL TEXT。
For example, we can send this to DeepSeek or any AI tool。
Highlights below can be adapted to your needs, and other direct reproductions can be applied。
WE'LL SEND AN AI TEST。
AI WILL FINALLY RETURN TO YOU WITH A SET OF HIGH-QUALITY HINTS, WHICH YOU JUST HAVE TO COPY。
And then paste to dream creation。
III. Elaborate editorial techniques
First move: Word design
THE TEXT IS THE WORST-HIT AREA OF THE AI BIOMAPPING, AND IT NEEDS TO BE BOTH CORRECT AND CLEAR IN THE FONT THAT AI PRODUCES, AND IT NEEDS TO BE WELL-DEFINED, IN THE FIRST PLACE AND IN THE SECOND PLACE
But to get a text style that fits your heart, it's a little bit of a trick。
I'm here to summarize the message:
- No less quotes
- Order is important
- It's big by the front
- Line Break quote
1. Quote not lessIncrease Text Generation Accuracy
THERE ARE TIMES WHEN WE WANT AI TO PRODUCE SOME EXACT WORDS ON THIS MAP, AND THERE'S ALWAYS SOME KIND OF GLITCH OR MISWRITTEN。
A GOOD SOLUTION AT THIS POINT IS TO WRAP IT UP WITH A DOUBLE QUOTE, SO THAT THE AI IS MORE ACCURATE。
for example:
A modern, dynamic spring ride poster with the words:Faster, life slows downIt's not like it's the same thing
The dream map also contains an icon with an automatic quote, as follows:
2. Order is important
The more important the text, the more important it is, the more important it is。
For example, we saw a poster case of this spring ride。
I gave it to you in the hint。
SO YOU LOOK AT AI AND YOU FOLLOW THIS SEQUENCE:
Now watch it. I'm not changing anything. I'm just changing the order of the main subheadings in the hints。
Top center: The purple background of Skycat International Logo。
LARGE MASTER TITLE (FLUORESCENT GREEN): "BIKE SEASON"
Yellow subtitle: “Most speed”, “Life slows down”, followed by “11.16”。
Bottom information (white small): Big words, " Lake thousands of islands / 09:30 – 13:30", followed by small words for multi-line details。
Look at this change in him。
See, now he's just turning this English phrase into our visual center font。
It's also the largest, so it proves that our order is the hierarchy that determines the text。
3. Breaking of lines and how to correct the text
For example, the text that you're going to generate on a picture is longer, and it's less than ideal to change lines. What do we do
It's simple, like this case, "Bring a squirrel to spring," the words. I didn't break it. I didn't break it with a quote。
And one of the things he can see on the picture is that all the words are in line。
I want such a vision. It's simple. You change the hints from a set of quotation marks to two. From "Take the squirrel to spring" to "Take the squirrel to spring"。
That's where the original quotation marks were put together, with two sets of quotation marks between, which is the effect below。
That's the key to getting you from being a kid to being a master。
Let's start with a simple reference。
Second move: Upload reference
1. Flier reference diagram
THAT'S, FOR EXAMPLE, HOW WE UPLOAD HERE A WAVE, AN IP IMAGE OF RABBOU。
And then let him help us with a new poster based on a reference map like this。
FILM LEVEL 3D RENDERING, C4D INTEGRATION TREND ILLUSTRATION
The center of the picture is a life-filled Labbu Tidal Toy
IP IMAGE。
The image is a very powerful low angle view。
# The petals and the starlight of the blurry future #
The center view of Rabbou floats on a geometrical platform that will feel so much in the future that it will not be able to reach the world
The background is a soft and gradual twilight sky。
We click on the upload reference and send it over。
And look at the poster that eventually came out, and it accurately identifies the image that we uploaded。
AND IT'S EXACTLY THE ONE-TO-ONE THAT KEEPS HIM IN THIS SHAPE, WHICH IS MORE ACCURATE THAN YOU USE 100 WORDS TO DESCRIBE A LONG EAR, HAIRY IP IMAGE LIKE THAT。
so pictures are more direct than words. look at the results. it's nice
And for example, you've already created this map, but you're not satisfied with one of the subjects in it. Like this one:
I'd like to replace this squirrel with another one, and I'd like to upload one of my reference diagrams, and the other one that I'm going to replace。
Directly orders:
Replace the image of the squirrel in figure I with the image of the squirrel in figure II to ensure color, shape, material consistency
let's see how it works. nice
This is the second case, where you've created a picture, and you want to replace one of the elements。
2. Upload multiple reference maps
Supports not only a single reference chart, but also a single multi-chart input, which supports complex editing such as grouping, migration, replacement, derivatives, etc., to achieve a difficult synthesis。
When uploading multiple reference maps, it is made clear that different elements for reference/editing of different pictures improve accuracy, such as:
I found the following three maps, which combine to produce city posters with text。
Broadly speaking: Add a large panda to the scene in figure 1 and can be found in the poster style in figure 3, and add a large panda in figure 2 as a well-crafted poster。
LET'S SEE WHAT WE CAN DO
Add the scene of figure 1 to the big panda, with reference to the image 3 poster style, to figure 2 big panda, to make it a well-crafted poster without output
Even the top right corner of the left map, the details of the Marseic。
If you want to change the details on some of the drawings, you can click into the picture and use the painting tool to adjust it。
It's much cleaner
We follow that line of thinking, and we're going to make this picture a beauty, a poster with text。
Now the picture is too short because of text。
LET'S START WITH A SCALE-UP, CLICK ON THE ONE YOU WANT TO EDIT, AND THEN CLICK ON THE SCALE-UP, WHICH AI CAN HELP US TO COMPLETE THE EXTENSION。
I've chosen the 9:16 ratio here because we need to vertically increase the height of the picture。
but i'd just like to add the above, and i'd like to keep the bottom on, and then i'd like to move this clipping box to the top. click to generate ok。
To look at the effect of the enlargement, the left side is before the extension, and the right side is after the extension, which is clearly larger。
And then we'll give the poster and the text, and you can find the corresponding text references. Then we'll use what we're talking about。
Look at the results
It's amazing, it's completely kept the style of words that we want. It's a good fit。
If you want to edit further, we can click on the canvas, and when you come in here, you can see that there are a lot of tools that you can use and that you can play with yourself。
I have expanded the following sections here, as well as the text and pen tools, with the following final results:
wind up
As you can see, we've just gone from three pieces of material to integration, and this combination of punches has been made, and an innovative poster has been created。
SO THE SET OF THINGS THAT WE'RE TALKING ABOUT TODAY IS MORE THAN JUST A SIMPLE AI PAINTING TECHNIQUE。
It's the bottom-of-the-board practical method. So, did you learn to work today