Artificial Intelligence (AI) is playing a significant role in everyday life. From perfecting Google searches to recommendations on Netflix and showing us related merchandize on Amazon. Photography is one sector which has been revolutionized by AI in recent years but has gone somewhat under the radar, changing consumer expectations.
Smartphones have become equipped with AI cameras that creates images that are better than ever whilst automated software is allowing professions to devote their time in being creative rather than focusing purely on taking pictures.
Here, we’ll take a brief look at the ways AI is changing photography and how technology is supporting that.
What comes to mind with AI?
When you hear the term AI, most people tend to visualize something of Terminator style proportions and immediately assume it is somewhere in the distance. However, it is important to realize that AI is all around us and refers to any machine or device that has an understanding akin to a human. A subset of AI is an application known as computer vision which has become prominent in accelerating the world of photography over the last few years.
Whenever a new smartphone is released, the public brace themselves for news of the camera quality. Today, the battle for the best camera doesn’t lie in the hardware, even though it’s still a necessity. It’s really in which brand can create the most innovative algorithms to provide the best shots possible. While it takes time and processing power to train these impressive algorithms, they can be run on low-powered devices without hassle. Both Google and Apple took many years to get where they are today.
AI is being used to augment and manipulate images to create quite astounding results.
Here’s our current AI situation:
The technology being delivered has become known as computational photography which is firmly driven by AI. Apple uses this to drive its portrait mode via the dual-camera phone. The algorithms can recognize people with one camera. The second camera works by creating a depth map to isolate the subject and blur the background. Whilst the technology associated with this isn’t necessarily new, doing it in real-time has been seen as a major breakthrough in photography. This is a brilliant direction where the future of AI in photography is headed.
Android and iOS are both offering impressive night mode features. These are built on AI that make images seem as if they are captured in daylight. Google and Apple have invested time in creating powerful chips to produce real-time photos that some dedicated cameras can’t offer. Deep Fusion on the iPhone uses AI to combine multiple pictures and create a single photo rendered with greater precision and detail. Technology truly is amazing! We’re at a point where AI delivers images and video at a level on par with professional equipment.
So what can we expect for the future?
With further development expected, AI tools may start to replace the technical skills required for photography. For example, Google Clips, the AI-powered camera, can automatically apply the right lighting and composition to an image. While reviews of the technology are far from spectacular, it tells us that Google’s moving toward automated photographers.
In 2020, AI-assisted image stabilization and subject tracking automation will be part of the critical feature upgrades of new smartphones. Right now, shaky hands and bounces can be quite visible with images and videos as people move. AI is helping smartphones cameras become increasingly capable of detecting multiple people and ensuring a sharp focus on the faces. Machine learning powered features allow users to draw a box around subjects. This helps to keep it in focus as your phone automatically rotates during a video.
Next generation smartphones will be able to differentiate between facial hair, clothing and objects to apply filters to each. This could apply effects and quality improvement at an object level as opposed to just the image in its entirety.
That said, Pixlr Editor will be releasing all-new AI integrated tools to help users (like you!) edit and modify your images better!
Generative Adversarial Networks (GANs) could also be on the horizon. This is a deep form of machine learning that can create photographs from text inputs. In theory, an entire video could soon be created just by feeding some code in GAN.
AI is having a major influence on how photos are edited and the technology required to do so. Humans are still required to add a creative and story-like element to the work. But a lot of what was previously manual labor could be sufficiently augmented with technology.