Extending the Attain of Your Pictures With AI

Everybody is speaking about AI.

AI (Artificial Intelligence) has discovered its manner onto our computer systems, our telephones, our on-line searches and our interactions with customer support. It appears as if each know-how firm now advertises “AI” as a part of their product. Indeed, some are portraying AI as a relentless companion, prepared to assist at a second’s discover with any of life’s challenges.

The latest explosion within the presence of AI betrays the truth that AI has been round for fairly a while and has been in improvement for a few years. But the comparatively latest introduction of “Generative AI”, a sort of synthetic intelligence that creates new content material—similar to textual content, pictures, music, or code—based mostly on patterns it has realized from current information, has put AI into the forefront of our digital interactions. The most well-known Generative AI platform is Chat GPT, though there’s a plethora of others in use at present.

Words & Photos by Michael Salvarezza & Christopher P. Weaver

Simultaneously, underwater photographers have seen the event of all kinds of cameras for capturing pictures of the marine world they’re encountering whereas diving. Whether it’s an costly DSLR in a devoted housing, a lighter-weight mirror-less digital camera, a degree and shoot system or an iPhone in a housing, at present’s divers can benefit from the course of of constructing and sharing pictures of their dive adventures extra simply than ever earlier than.

But how are you going to make higher use of the lots of, maybe 1000’s, of pictures you’re creating underwater? How are you able to distinguish your work from the work of so many different divers who’re additionally posting their newest pictures to social media every day?

In brief, can you utilize AI to increase the attain of your images and entice the eye of various audiences from the standard one used to seeing underwater pictures?

Since Generative AI has the flexibility to create new content material, it may be used to create model new pictures based mostly on instructions (or prompts) given to it by a consumer. We needed to discover a special angle, although, by asking AI to switch our current underwater pictures.

Suppose, for instance, that you simply had a shocking picture of a sea turtle. The authentic we included on this article was a loggerhead turtle photographed on a reef in Aruba. What when you needed to share this with youngsters with the aim of fascinating them in marine life? Everyone is aware of that cartoons are sometimes used to attach with youngsters, so what if we used AI to rework the Loggerhead Sea Turtle right into a cartoon? We used essentially the most primary stage of ChatGPT for the needs of illustrating this text and utilizing a easy immediate that learn “Turn this picture into a cartoon” we created the cartoon model of the turtle. Instantly, we now have discovered a brand new use of our authentic picture. Indeed, we may create a whole youngsters’s ebook and illustrate it with cartoon variations of our images.

If we needed to create a scarier model of our Sea Turtle, we will instruct ChatGPT to change the picture to be extra menacing. The ensuing picture may cross as a villain within the story in a youngsters’s ebook!

Maybe the viewers you are attempting to achieve isn’t youngsters however folks keen on tremendous artwork and work. Using ChatGPT, we rendered the identical Sea Turtle in two completely different creative kinds: Impressionist and Van Gough fashion. The ensuing pictures are a very completely different re-imagining of our Loggerhead Sea Turtle. They would make for fascinating wall hangings, and maybe there are others who may additionally discover them to be compelling as artwork work.

An picture we created of a Porcupine Pufferfish within the Sea of Cortez was simply remodeled into an expressionist fashion portray. Exploring additional, we requested ChatGPT to render this identical picture as if Salvador Dali had painted it: the ensuing surreal picture represents yet one more manner of taking a look at this intriguing fish.

And a really good picture of Anemonefish taken on a reef on the southern atolls of the Maldives is remade as a portray completed within the fashion generally known as “realism”. Both the unique {photograph} and the brand new AI rendering are visually intriguing.

The great thing about utilizing Generative AI on this manner is that the software program is extraordinarily subtle and is consistently studying from the large quantities of knowledge on the web, that means that the makes use of of this software program to increase the attain of your images is restricted solely by your creativeness.

For instance, perhaps diving is of nice curiosity to you however to not your teenage youngster who’s extra keen on video games, comics and superheroes. Trying to curiosity them within the underwater world might require a special strategy. Using ChatGPT, we created a superhero themed model of a diver exploring the crystal-clear waters of Silfra in Iceland.

We even needed to have a little bit of enjoyable and used a picture of a shrimp photographed in Lembeh, Indonesia because the mannequin for a film poster designed as a Nineteen Fifties-era Science Fiction film. The outcome from ChatGPT, which created the fictional film title and inserted the screaming actress by itself, is amusing!

Finding methods to achieve completely different audiences together with your images is difficult. Using AI to re-envision how our pictures are seen by others might open a brand new world of potentialities for underwater photographers.

What about digital post-processing? Every underwater photographer is aware of that the majority instances the unique picture captured by the digital camera wants some “cleanup” on the pc to boost the image or take away imperfections. Software applications similar to Photoshop are the gold requirements for this kind of work, however they are often costly and complex. Can Generative AI carry out the identical features because the digital postprocessing software program?

The brief reply is “maybe” and “it depends”.

Generative AI typically does precisely what it’s designed to do: generate new content material. This is vital as we glance to scrub up our pictures. For the needs of this text, we took an authentic picture of a Goliath Grouper photographed on a shipwreck off Destin, Florida and requested ChatGPT to scrub it up. The particular instruction we used was to “clean up the image by adjusting the colors, sharpening the focus and removing backscatter”. As you’ll be able to see, the AI generated a really good rendering of the Goliath Grouper in consequence.

Words & Photos by Michael Salvarezza & Christopher P. Weaver

But watch out: look very intently and also you’ll discover the spot patterns on the fish are completely different from the unique, the attention appears to be like completely different and the enamel in its mouth have been altered. Even a small starfish clinging to the roof of the wreck has been modified.

The warning we’re elevating right here is that the cleaned-up picture of the Goliath Grouper is good however it isn’t the unique fish. If you have an interest in accuracy, or are utilizing your images to doc marine life (for instance), the ensuing pictures from AI post-processing is probably not what you’re in search of and is probably not legitimate or applicable on your functions. Proceed with warning!

Finally, it’s important that every one photographers take note of the ethics of all of those new capabilities. As new know-how emerges, new authorized and moral questions come up: Who is the “author” of content material created by Generative AI platforms? Can these ensuing pictures be copyrighted? What obligations do photographers need to disclose that AI was used within the altering of a picture?

For instance, when creating “artistic” variations of your authentic images, present U.S. Copyright regulation states that when you created it your self (e.g., you used an AI instrument to render your picture as a portray), you usually personal the rights to that picture, so long as the AI instrument’s phrases of service provide you with these rights. Most main AI artwork instruments (together with ChatGPT) state of their phrases that you personal the output you create — together with when it’s based mostly by yourself picture. In the U.S. (and most international locations), AI-generated works alone can’t be copyrighted by the AI, but when you make significant artistic selections — like deciding on the fashion (e.g., “realism,” “Van Gogh”, and many others.), composition, prompts, edits, and many others. — that human involvement can provide you copyright within the ensuing picture.

We consider that the moral panorama of utilizing these platforms is quickly evolving. In some ways, at present’s software program is akin to the canvas and paintbrushes utilized by conventional artists. As such, and as acknowledged above, so long as the unique picture the ensuing artwork is predicated on is yours, you need to “own” the computer-generated rendering. But utilizing another person’s materials as a supply with out their permission is clearly unethical, even whether it is being remodeled into a creative model. And in all instances, photographers ought to disclose when their pictures have been adjusted by AI earlier than they’re submitted to be used apart from for private functions (e.g., journal work, scientific information, and many others.).

Generative AI opens a completely new and just about limitless world of chance for the usage of your underwater images. Extending the attain of your pictures could be rewarding and significant, particularly if you’ll be able to encourage others to care about our stunning, and fragile, underwater world.

And apart from, it’s simply enjoyable!

Words & Photos by Michael Salvarezza & Christopher P. Weaver