It is no longer a problem today that images change into text.
Various OCR functions make it easier for you to extract text from pictures.
It is not a big problem to let AI explain a picture.
However, drawing is still difficult for AI today, and identifying image extraction information is processing information for AI.
However, drawing is a new layer, which not only needs to process information, but also needs to complete creation.
The former is a multiple choice question, and the latter is a proposition composition.
Only after the multiple choice questions are answered well, the next step is to answer the subjective questions that can be played freely.
But nobody thought that TikTok was the first person to answer this question on his own app.
Using TikTok to generate Facebook, Ma Yun and other giants such as Google and OpenAI who have invested a lot in the AI industry for a long time, TikTok may just be a "transition student".
But why is it not surprising that the shift change student first made a difficult problem by bypassing the giant? Although the shift change students have also made some ingenious efforts, at least it is still very impressive.
TikTok's text to image function is called "AI Greenscreen", which replaces the monotonous white background and allows AI to generate the video background for you.
The background of these videos may not meet the needs of the creators, but after all, this is a unique green screen background customized for you.
If you are lucky, it may also match the theme of the video content better.
Click the special effect to experience this function when sending video The picture comes from the Silicon Star people, so we did some tests with this new function to see what the picture TikTok drew is like.
Among the words in these random tests, some of the paintings were presented in a confusing way, while others were rated as "well depicting the weird feeling".
What is highly praised is the painting presented by entering "Facebook".
You can vaguely recognize the blue icon of Facebook.
The single eye and one ear in the picture create a unique sense of horror.
Combined with the recent news of Facebook, I have to say that this painting depicts the feeling of the word Facebook very well.
There are also words like Apple and China that are accurately depicted in the image.
The former can easily see that it is an apple, while the latter can also see Chinese style buildings, while avoiding such sensitive patterns in the field of second innovation as the national flag.
It's just hard to show the image of Apple, even if the key words of Apple+Apple CEO Cook show the image of Apple and food.
However, there are many irrelevant ones.
For example, we tried Alibaba, Tencent and ByteDance, and the generated ones cannot be said to have nothing to do with these brands.
However, it is difficult to recognize at a glance that it is somewhat abstract.
There are many interesting pictures drawn by inputting the names of people.
The name of Li Ziqi, a famous manual blogger overseas, creates a comfortable landscape painting;
The pictures produced by the famous former Premier League coach Arsene Wenger can also be recognized at a glance, which are classic photos

give likes(0) Reward

Comment list 共有 0 条评论