In the past several years automated intelligence and machine learning has become increasingly more accessible to every day users. In the past, this type of technology was limited to research laboratories, universities, and fortune 500 companies, but is now completely free to virtually everybody.
If you are a content creator, you can leverage the power of AI by using a plug-in called ClassifAI, which allows you to automatically tag and categorize images and content in WordPress.
Watch the video above, or read through the automatically generated transcription to get a good understanding and idea of the power behind this tool and how it works.
Hey, guys, this is James with Isotropic. And in this video, I want to talk about how we can use artificial intelligence to automatically tag and add alternative text to our images in WordPress. So this is a really neat plug in that I was introduced to by a client. It's called ClassifAI. And essentially what it does is allow you to connect your WordPress website to machine learning and artificial intelligence from IBM and Microsoft and then take that that machine learning power that you have access to with these services and automatically tag an image, give them image descriptions and crop them at the focal point.
You can also set up tags. So if you have a massive image library, adding the image tags makes it easier to search. So this is this is really made for publishers. If you're using WordPress to publish a lot of content and want to kind of slow things down and speed things up and add automatic alternative text and a title and a caption to your image, then that is this this could be a really good option for you. So I want to show you kind of how it works on the front end (ish) and then show you how it works in the back end.
So I'm first on Unsplash and I'm going to find a photo of some recognizable object, and then I'm going to also use something a little more difficult just to kind of show you what this can do. So let's just look up a camera. And if you don't know, Unsplash is a great photo repository, completely royalty free images, so here's an image of a canon camera and we all know this is a camera. But I want to show you kind of how the machine processes it and what it does.
So we're going to upload our media like normal. Drag and drop into WordPress to upload it. I have an image optimization plugin, so it will take a little bit to crunch it down. And once it's crunched down, I'm going to show you the tags that were assigned to the image, the alternative text that was assigned to the image. And if we're lucky, we might even get a title and a caption out of this. And as you can tell, if you're uploading a million different images at once, if you're doing a lot of content every day, this can really help you save a lot of time and remain in good terms with search engines and accessibility and all of that.
So here's our image uploaded. And you can see we have ClassifAI image processing. That's a plugin that we installed. And I'll show you how to do that in a bit and. Here's the alternative text, a camera on a table, so it was able to figure out that this is a camera and it was able to figure out that this is a table. And then it also tagged everything, tons of really good tags. And then it also gave a description, Canon EOS.
So this is this is incredibly impressive to me because. It might have been difficult for the machine to figure out that this was a Cannon camera because of the reflections and it's a dark object, but he figured out it was a camera. It figured out that the camera's on the table. It figured out that it's a cannon camera. And then it went as far as adding digital SLR, mirrorless, interchangeable lens, camera point and shoot camera. These are all different types of cameras that you might want your image to be classified.
But let's do another image just really quicklly iron that point home and let's look up, I want to look up a lion and see if it can figure out living things. So we're going to download a picture of a majestic lion, and I'm actually going to do a smaller file size if we go to that drop down menu and click small, smaller file size, less space used up on your site. And it usually looks pretty good on the front end, even in that size.
So we're going to upload this. It's going to crunch - and the Azure service will run and it will automatically tag the image using A.I. So let's take a look and see what see what this one did. And this is actually live on our our agency website now, animal, carnivore, flying, mammal, outdoor. A lion sitting in a field. And. What do you know, it's a Lion and it's in a field. It's an animal and a carnivore and a mammal.
So that's really impressive to me, and I'm very surprised that machine learning is able to do this and that we are able to access this for free. So now let's talk about the plug in behind this and how it assigns these terms to it. So plugin is called ClassifAI, as I pointed out in the beginning of this video, and I don't believe it's available on the repository, but it's a free plug in. So what you need to do to download it is add your information here at a use case and then register for free.
And once you've registered for free, you'll get an email and that email will go ahead and send you a registration. Registration is 100 percent free, provides notifications, all of that. It'll send you an email. So you add your registered email and your your your registration key and then save the changes in the ClassifAI plugin. And this plugin actually has two different aspects. It has language processing, which I'm not talking about in this video. And language processing allows you to essentially access the advanced tag management for your your your content.
So it will automatically tag and categorize content on a website. I'm not sure how much I would want to use that because I like tagging and categorizing my blog posts on the isotropic blog manually. But I could see how some people want that. Now, if we go to image processing, it'll show you. How this works, so. First off, we use Microsoft Azure. Or however you pronounce it, I hope I pronounced it right, and this is actually using a sub offering of the Microsoft Azure platform, which is computer vision, and you need to set this up on the Azure platform.
And once you've set it up and a link in the description to a guide showing you how to do that, basically you go to the platform, you register for computer vision, you get twelve months free. I think the free plan you get. Maybe 20 API calls a minute, which means you can upload probably 10 images, I don't I don't really know how it works precisely, but you get you get a decent amount of free stuff for 12 months.
All you need to do is register for this platform and then enable the computer vision for the marketplace. Setup the endpoint, grab the first API, copy, paste it in, and you're good to go. And what this does is connect it to computer vision. Computer vision is a tool that Microsoft made, which basically does exactly what I just showed you. It it looks at images and it recognizes the content within the image and then it automatically assigns language to that image.
So as all text or and as tags. All right, so here are the options that we have a fairly basic options, but it's it's does all right automatically caption image and caption confidence threshold is important. So this is 75 percent sure that the item is what it is. If it's under 75 percent sure, you will not get a caption, automatically tag the image and then that's the threshold for that. And then image tags. If you have a custom taxonomy, I think you would be able to assign it to that enable image smart cropping.
If there's a region of interest and the the computer determines that the the rest of the image isn't going to be beneficial to the readers because it's a bunch of dead space, it will crop it to the subject just as they kind of show you here. This is your big image. If you crop it down to a thumbnail in the default version of WordPress, it will. It may and now will and may cut off the actual the actual subject, but what this does is uses the computer vision to determine where the center mass of the main focal point is and then cropped that square thumbnail around it.
So it really helps to get better thumbnails. And that's a big selling point and it's done automatically. But if we run back here, you got smart hopping in, the final thing you have is OCR and the OCR isn't great and let's see if I can find. A neon sign and see if we can show you. How the OCR works, it isn't great and it might be something that you don't want or we'll do this, we'll do it one way, son, it might not be something you want because it's not always.
Very, very functional, for example, I uploaded a picture of an Audi car and that Audi car, it grabbed the text that said Audi from the front of the car except the D in the Audi became a Russian character rather than a D, so I don't know how beneficial that would be for search and it might just confuse your visitors. But for block letters, I think OCR should work. All right. And I'll show you how it how it works here.
First off, super impressive. A close up of a street sign sitting on the side of a road street sign. It figured out that this was a road. It figured out they were kind of close up to it, probably based on the focus and the background. And that's really impressive. Here's our OCR. The OCR gets pushed into the description and it gets one way. And then traffic sign, text street, Sky Road, all of that.
So as you can see, this is an incredibly powerful little utility that not many people know about. It's the ClassifAI plug in. And this is just the just your automatic image stuff. I'll go over language processing in another video once I have that set up. But this is just the automatic image stuff. And essentially what it does is it automatically adds alternative text that's really descriptive and is is surrounding aware. So it knows that this is a line, but it also knows that it's in a field.
It will tag your images, which makes it easier on the back end if you're trying to search for things in the future. If there is text to be translated and you have OCR on, it will add that into the description box. And finally for the thumbnail crop.
It will go ahead and crop around here instead of maybe the default being this crop, so you can just see that this is a really powerful plug in, allows you to access some really powerful technology that somebody wouldn't even have thought would exist 10 years ago. Just a really cool thing that I wanted to introduce to you guys and kind of show you it's free to install. So go ahead and install and poke around and be wowed and amazed. Like me, I've probably uploaded a gigabyte of images from one splash just to see what the machine learning and AI spits out.
So it's kind of addicting and it's just a really interesting and neat utility. All right. This is James from Isotropic. If you liked it content, give us a like comment. Subscribe also head over to our well followed blog. We do a lot of WordPress tutorials. There's probably going to be an even paper tutorial on how to install your your API key, an endpoint you URL because I'm going to link in the GitHub to the instructions on how to do that, but it's a little bit complicated.
So I might make a video or a blog post and link it in the description showing you how to do that. All right. Have a great day.
IsoGroup- Web Dev/Design, WordPress and More