How has artificial intelligence been changing mobile photography?

The most recent advances of mobile photography have come from AI, in addition to the superior technology of sensors and lenses.

If you’re wondering how well the next mobile camera will work, pay attention to what the manufacturer often says about artificial intelligence (AI) – technology that makes incredible progress. of photography village over the past few years, and there is no reason to think that this process will slow down, technology experts The Verge said.

In fact, besides the new technologies from sensors or lenses, it is clear that the most recent impressive advances in photography also come from software and processor chips. All of which creates a complete AI system, which helps the camera better understand what we’re looking at and wants to capture.

Picture Capability The super-impressive moon of the Huawei P30 Pro is a testament to the practical value of what the AI ​​delivers. Perhaps mobile photography technology has “taken a step to the cloud” has made users feel overwhelmed but at the same time not yet understand what is happening, leading to skepticism and “backward”.

The image of the Moon is captured with the Huawei P30 Pro’s periscope lens in 50X zoom mode.

According to Huawei ‘s description, the Moon Mode (Moon Mode) also works on the same principles as other AI modes, helping to recognize and optimize the details in the image so that users can capture the pictures. Better photos. AI mode cannot replace the image being taken with another image, because it is unimaginable to require a huge amount of storage.

AI mode on the Huawei P30 Pro can identify more than 1,300 objects and landscapes. Based on machine learning principles, the camera recognizes an object and landscape, helps optimize focus and exposure to enhance the details of the object and landscape such as shapes, colors and points. light / weak light, …

Also note that AI shooting can be turned on or off easily while taking photos. Users can take photos of the Moon with the Huawei P30 Pro through periscope lenses without turning on AI mode.

Another AI-related story is Oppo’s AI Beauty selfie technology, such as Oppo R11 Pro. This smart beauty mode can identify 8 million unique features, suitable for different groups of users worldwide based on machine learning technology, can even identify up to 4 characters in the same Selfie photo frame. Even before shooting, users can manually adjust the features: face, jaw, eyes, nose, skin … to create the best photo. After many years of development, it is clear that AI Beauty has helped the faces of male and female celebrities become more beautiful when they are on the screen.

AI from computer software posted on smartphones is creating many breakthroughs.

Back four years ago, when the Google Photos app launched in 2015, Google brought a clear demonstration of the powerful combination of AI and photography. Previously, Google used computer learning to classify images in Google+ for many years, so the launch of Photos app with AI technology was aimed at features that most people could not think of. The result is a user’s haphazard photo library of thousands of unedited photos that are organized into a searchable scientific database.

Google has built a technology-based AI system from DNNresearch that they purchased in 2013, by establishing a network of deep neuron-trained neurons. This is called supervised learning: This process involves training the ability to learn on millions of images to be able to search for visual clues at the pixel level, thereby helping to identify categories. for each photo.

It takes a lot of time and processing power to create an algorithm like this. After the data centers have completed their work, it can run on mobile devices without any trouble. About a year after Google Photos was launched, Apple announced a similarly trained image search feature.

AI can identify the scene before making appropriate settings.

Smart photo management software is one thing, AI and machine learning are thought to have a greater impact on the way of original photography. Of course, the lens continues to grow a bit and the sensors may be slightly larger, as Huawei overcame physical limitations when cramming the 5X zoom optical system into the ultra-slim P30 Pro.

As a result, today’s phones take better pictures in some cases when compared to many dedicated cameras, at least before post-processing. That’s because traditional cameras cannot compete with another type of hardware – the system chip contains CPU, image signal processor and a neuron neuron processor (NPU).

“To be honest, out of all the possibilities and applications offered by the AI ​​wave over the past few years, photography is the most realistic field. Camera is an essential feature of any phone, and AI is the best tool to improve it, ” said The Verge expert.

Source : 24h