How AI Technology is Changing Photography and How Canon Cameras are Incorporating AI

Canon Eos R3 Review Canon Cameras Eos R1

I thought I could write a short post about artificial intelligence and what it means for photography. Nowadays everyone is talking about AI technology, often mystifying it. Here are my 2 cents, with an eye on Canon cameras, obviously ;-)

Introduction

Artificial intelligence (AI) is changing the world of photography in many ways. From the emergence of new text-based diffusion models that allow anyone to create an image by typing a description, to the launch of more powerful AI image generators like DALL-E 2, the development of open-source tools like Stable Diffusion, and the experimentation of photographers with AI-generated creations, AI is transforming the way we think about photography.

Big players like Google and Microsoft are also developing more powerful AI tools. However, ethical and legal questions remain, and AI-generated images could start to replace photographs in specific areas like lifestyle stock imagery.

Canon cameras and AI

Canon cameras are incorporating AI technology. For example, the Canon EOS R3 features an AI-based autofocus system that can track subjects with incredible accuracy. The camera also has an AI-based image processor that can fix most major digital photo issues, such as distortion, vignetting, and chromatic aberration. AI is changing photography and Canon cameras are at the forefront of this transformation. As AI technology continues to evolve, we can expect to see even more exciting developments in the world of photography.

Canon’s Deep Learning AF

I decided to give this topic the most space, since it’s foundational to Canon’s approach to AI for auto-focus. To learn more about Deep Learning you might start here.

Canon’s Deep Learning AF is just one example of how AI is changing the photography industry. Deep Learning AF can improve the accuracy and speed of autofocus, which can help photographers capture better images. The technology can also help cameras focus on subjects more quickly and accurately, even in low light conditions. The algorithm was trained using Canon’s image database of sports photography and pictures of the company’s ambassadors who shoot sports. This has resulted in a highly accurate autofocus system that can track subjects even when they are moving quickly.

Deep Learning AF uses machine learning algorithms to analyze images and predict where the camera should focus. The algorithms are trained on large datasets of images to improve their accuracy over time. The technology can also learn from user feedback to improve its performance. Canon has integrated Deep Learning AF into several of its cameras, including the EOS-1D X Mark III and the EOS R5. The technology has enabled features such as Eye Detection AF and Head Detection AF, which can help photographers capture sharp images of people.

Canon plans to continue developing Deep Learning AF and integrating it into future products. By using machine learning algorithms to analyze images and predict where the camera should focus, the technology can help photographers capture better images. As Canon continues to develop and integrate Deep Learning AF into its products, we can expect to see even more advanced autofocus capabilities in the future.

Image Enhancement

AI is also being used to enhance the quality of photographs. By leveraging deep learning algorithms and neural networks, AI can improve various aspects of an image, even on the camera itself, including resolution, color balance, noise reduction, and more. Some of the key ways AI assists in image enhancement are:

  1. Super-resolution: AI algorithms can upscale low-resolution images by reconstructing high-frequency details and textures. This process, known as super-resolution, generates a higher-resolution image without sacrificing quality.
  2. Noise reduction: AI can effectively reduce image noise caused by low-light conditions or high ISO settings. By analyzing and differentiating between noise patterns and image details, AI can suppress noise while preserving essential information.
  3. High dynamic range (HDR): AI can help merge multiple exposures of the same scene into a single image with an extended dynamic range, preserving details in both the shadows and highlights. This creates a more balanced image with increased details and richer colors.
  4. Color correction and white balance: AI algorithms can analyze an image and make automatic adjustments to color temperature, saturation, and contrast, resulting in more accurate colors and improved overall image quality.
  5. Object and scene recognition: AI can identify and recognize objects, scenes, and even people within an image, enabling cameras to automatically adjust settings and apply appropriate image enhancements based on the subject matter.
  6. Autofocus and stabilization: AI can assist with faster and more accurate autofocus (see above), as well as image stabilization, by predicting subject motion and compensating for camera shake in real-time.

Conclusion

AI is changing the photography industry in many different ways, and it’s just the beginning. The best has yet to come. From improving autofocus systems to enhancing the quality of photographs and automating the editing process, AI is making it easier and more efficient to take great photographs. As AI continues to evolve, we can expect to see even more exciting developments in the world of photography. Stay tuned.

Save big on Topaz Video AI – Get It For $249 (reg. $299)

Topaz Video Ai

You can save $50 on the new Topaz Video AI. Read on to learn more.

At a glance:

  • Fewer motion artifacts
    • Other video upscaling techniques often create a “shimmering” or “flickering” effect from different processing in adjacent frames.
  • Recover video details
    • Inject real details into your videos, derived from the additional information in multiple adjacent frames.
  • Get more natural results
    • Traditional upscaling often causes artifacts. We mitigate these by deriving new information from multiple frames.
  • Trained on videos
    • Topaz Video AI was trained exclusively on video clips instead of still images.

Until March 31, 2023, you can get Topaz Video AI on sale at $249. Compare at $299.

Production-grade AI models for professional use cases

Topaz Video AI focuses solely on completing a few video enhancement tasks really well: deinterlacing, upscaling, and motion interpolation. We’ve taken five years to craft AI models robust enough for natural results on real-world footage.

Topaz Video AI will also take full advantage of your modern workstation, as we partner directly with hardware manufacturers to optimize processing times. (Many of them already use Topaz Video AI to benchmark AI inference.) Own the software and use it for as many projects as you like, right in your existing workflow.

Topaz Holiday Deal: Save 25% On All Apps, 30% On Image Quality Bundle

Topaz Labs

Topaz Labs started a new rather interesting promotion for the festivities.

Until 12/31/2022 you get 25% discount on all Topaz Labs applications and 30% discount on the Image Quality bundle. Checkout the link and scroll down to see all offerings.

More selected deals:

Canon Releases Photo Culling App With Artificial Intelligence (iOS only so far)

Photo Culling

Canon has released an app for the most indecisive photographers out there: the new Canon Photo Culling app. 

Currently available on Apple iOS, the Canon Photo Culling app uses Canon’s new AI known as PHIL (Photography Intelligence Learning) to help determine your best photos based on scores against four models: sharpnessnoiseemotions and closed eyes. The app will also suggest deleting duplicative photos. Whether you’re looking to trim down your photo gallery to save precious space on your phone or just trying to find the best of the best, Canon’s Photo Culling app will make your photo management a better experience.

Available for download via the App Store, the Photo Culling app has two subscription options: $2.99 per month or $14.99 for the year, each with a three-day trial. For more information, please visit, usa.canon.com/photo-culling-app.

Canon press release:

Need a Digital Photo Assistant? Canon U.S.A. has you Covered with the Company’s First-Ever Photo Culling App

Canon Also Publicly Debuts “PHIL,” the Company’s Computer-Vision Artificial Intelligence Engine

MELVILLE, NY, February 11, 2021According to a recent report from Keypoint Intelligence, 1.4 trillion photos were taken worldwide in 2020 and 7.4 trillion images were stored1. With these staggering numbers, it could be overwhelming for the average consumer to decide what photos are best to keep and store. Delivering expert photo advice in the form of an app, Canon U.S.A., Inc., a leader in digital imaging solutions, today announces its Photo Culling app. Compatible with iOS devices2, this brand-new app helps determine a user’s best photos based on scores against four models: sharpness, noise, emotions and closed eyes, using Canon’s Computer Vision Artificial Intelligence Engine known as “PHIL” (abbreviated for Photography Intelligence Learning) 3. Additionally, the app will suggest deleting duplicative photos, making it easy for the consumer to free up space on their smartphone.

“In today’s ever-changing and overwhelming world, where thousands of photos are captured and stored in a person’s smartphone, consumers need an expert, reliable and intuitive photo tool to help them decide the best photos based on years of trusted knowledge and technology,” said Tatsuro “Tony” Kano, executive vice president and general manager of the Imaging Technologies & Communications Group of Canon U.S.A., Inc. “Canon U.S.A.’s new Photo Culling App is the answer and we are proud to see how the company’s Computer Vision technology within this app can assist consumers with finding and keeping their best photos of their fondest moments.”

The Photo Culling app features two culling options:

  • Whole Culling: Determines the best photos based on the absolute best scores against the four models: sharpness, noise, emotions and closed eyes. If the score is over the threshold the user sets, the photo is considered to be the best. The remaining images would be suggested as deletions.
  • Similar Culling: Determines the best photos based on score comparison among a similar group of photos. The highest-scoring photos within the group are considered the best photos. A second photo may also be selected as the second-best photo within the group. The remaining images would be suggested as deletions. The app is also able to find similar photos and group them together. For example, If a user selects 10 photos of a dog and 10 photos of a sunset, it will break the photos into two groups and find the best photo of each; one of the dog and one of the sunset.

In addition to its culling options, the Photo Culling app also features:

  • Photo Counts and Storage: Shows the number of photos a user has as well as the amount of phone storage on the Home Screen
  • Event Album: Categorizing and placing photos in albums on the Home Screen. These are dynamic albums based on events/date throughout the year and the app will pull events that have a large number of photos and suggest the user review these photos to delete and save space
  • Features in the Setting Screens: Displaying the number of photos deleted, the current subscription plan (see details below), option for notifications, parameter setting for each score against sharpness, noise, emotion and closed eyes. Users also can adjust these settings, such as changing how much of an impact noise has on a photo versus closed eyes4
  • Dark and Light Mode Settings Screen Option: Displays screens either as a dark or light mode option

More About Canon’s AI-Powered Camera, The PowerShot Pick (price below $400)

Canon Powershot Pick

More details about the Canon PowerShot Pick, a camera with artificial intelligence on board, emerged on the crowdfunding page.

It seems the Canon PowerShot Pick is having a huge success on the Makuake crowdfunding site. They already surpassed the goal by 5,900%. This is how the PowerShot Pick is introduced (machine translated):

Bringing together Canon’s advanced technologies! With optical technology and video processing technology, the camera itself makes smart decisions and automatically takes pictures and videos.

The small and cute design blends into the family circle and automatically records “memories of the whole family” with “natural and rich facial expressions”

The app automatically selects recommended photos and videos. It’s also fun to find the “best shot taken by chance”!

Canon touts the ability of the camera to determine when it is best to snap a picture:

With Canon’s image processing technology, PowerShot PICK itself automatically finds a person’s face and notices changes in the image reflected on the lens, so it automatically determines “now is the photo opportunity” and shoots.

The camera is meant to just stay there, without being noticed, and to take “memories” at the best time:

Small and light, this camera is designed to blend in with your family so that you can forget about it, so you can stay close to your family at home or out.

The Canon PowerShot Pick…

[…] recognizes the faces of people around you and automatically tracks them. The angle of view is automatically adjusted while making full use of pan, tilt, and zoom based on face information such as the number / position / orientation / size of the face using a unique algorithm while observing the state of the subject.

In addition, the characters shown in PowerShot PICK will be registered automatically. If you want to shoot more specific people, you can add a favorite mark to that person to automatically shoot photos and videos with priority.

The Canon PowerShot Pick understands voice commands:

The price in Japan for the Canon PowerShot Pick is 42,900 yen, which amounts to approximate $390. At this price it might become a huge success for Canon.

And here is an overview:

[via PetaPixel]