Very good to everyone.
In cyber, as in many other fields, we need to have good access to banks of images, videos and other visual resources with which to enrich our presentations or day-to-day reports, among other documents worth highlighting. In today's post I wanted to compile some AI-based tools that are becoming popular lately for the generation of visual content. There are so many projects that it is difficult to try all the ones that sound most interesting, so I have taken the opportunity to share with you some of the ones that are working best for us (plus a surprise that I have left for last :P).
I'll start with one of the largest, the
Microsoft's own Bing project, which is based on Dall-E. It is accessed through Bing, and allows us to generate images in a matter of seconds. For example, I have generated a series of photographs of a lynx jumping over a padlock:
It looks like a very marked style if we don't refine the prompt a bit more, but for a first generation the result is amazing.
The second one I will tell you about, and probably one of the main ones today, is
midjourney.com, which with that same console input would generate something like the following in a first iteration:
We use Midjourney as a professional tool, and it saves us a lot of work when generating visual resources for presentations. The results are truly incredible and for a really low monthly cost (for professional use).
The third one that I will list is
playgroundai.com. It does not achieve as good results as Midjourney, or, at least, I have not been able to get the same benefit from it, but it allows you to create reasonably well-constructed images at no cost, and, yes, it is more versatile than other projects, allowing you to upload images in which get inspired, play with size, exclude parts, play with quality/detail, etc.
The fourth project that I wanted to show you in this list is
pixlr.com, a project that I personally really like to generate illustrations. For example, below I show you an illustration VS a realistic image generated from the same data input:
The style of the illustration is very good, and by polishing it slightly it can become professional content.
The fifth AI that I wanted to show you today is
craiyon.com. This AI generates very good quality images. However, the AI that interprets the data entry is not the best of all, and to obtain the expected result you need to stick with the prompt for a while, as you can see below (if you see the taxi somewhere... I you say :P):
The sixth project I wanted to show you is
dreamstudio.ai. I like the quality of the images generated... but the result obtained does not fit perfectly with what was requested.
And you will see it quickly if we compare it with the results obtained with that same entry in Midjourney, which seems to be the only one that "understands" the meaning of "drive":
You can animate all these images with
convert.leiapix.com. To show you a simple example, I have animated the image of the guy that decorates the top of our blog (by the way, generated with Midjourney). It seems to come to life...
And for those who want to generate visual content but with a more advertising purpose, other types of projects such as flair.ai may be useful. As a quick summary, it is an online tool for building compositions to advertise products. It is very focused on stores and manufacturers, and works especially well with realistic content.
Of the 8 that we have seen, we would stick with Midjourney and Bing, but the key to all these AIs is undoubtedly "knowing how to ask", logically asking for what we want to design (better, in English), setting a visual style, using other images that we like as a base, and applying variations, such as, for example, eliminating a certain object. Working with AIs is really simple, but achieving professional results that require a little retouching to close them requires some practice.
P.S. I didn't want to say goodbye without sharing 2 more things with you. First of all, a last AI as a combo, because... there is no self-respecting hacker talk without its corresponding memes (if not, Daniel González and Elías Grande were not going to forgive me... :P). The website supermeme.ai will allow you to generate memes in image or gif, from more than 100 languages and in a matter of seconds:
With supermeme you will no longer need to save every meme you find on the Internet on Telegram..., and Dani and Elías will no longer have an excuse to prepare that meme talk that they always say that they are going to do and they never do... :)
And, as a second combo, and although it is not exactly a project based on AI, I leave you the link to futureailab.com, a portal where they are maintaining a index of AI-based applications, in which you can apply filters by topic to locate what you are looking for:
I hope you liked the compilation! From time to time we will share with you new projects that sound interesting and that you can make good use of in your daily life.
Greetings!
