The Dawn Of Intelligent Camera Assistants
One of the earliest products to feature AI in cameras, is the Arsenal intelligent camera assistant. It is not a new type of camera, but a device that attaches to cameras for added functionality. What Arsenal does is help photographers of any skill level get the best in terms of exposure and quality from their existing digital camera (e.g. list of supported cameras). The assistant uses AI to suggest settings that are based on your subject and environment. With advanced neural networks (NN), optimal settings for a scene are selected for you. The algorithms work similarly to decision making in self driving cars, but of course not as risky. You are operating a camera with this device and not a vehicle.
What Arsenal provides to DSLR cameras is already being experienced by smartphone camera users. That is because most flagship brands (e.g. Galaxy Note 10+, P30 Pro, iPhone 11 Pro Max) use plenty of AI already when taking photos. The addition of an AI chip, like in Apple’s A13 Bionic chipset called the “Neural Engine” provides the processing power needed to a smartphone camera to take better pictures. It doesn’t require a professional and it is done without any post-editing or retouching required after taking the shot. It is all done by the smartphone camera software before, during and after the image is created.
How It Works
The device attaches to the hot shoe mount on the camera. Then it plugs into the camera using a Micro USB cable. The device also has Wi-Fi capabilities to connect to the Internet to download firmware updates.
The intelligent camera assistant determines the camera settings for you and all you have to do is take the shot. The assistant comes with an app that is installed on your smartphone. It can then be controlled and accessed from there, providing a convenient way to wirelessly control the device. The communications between the device and smartphone has a range of 100 feet.
From Arsenal’s own website, it describes the process:
- Arsenal quickly examines the scene. It uses image recognition to identify environment and subject-specific needs (e.g. fast shutter for birds or camera vibration).
- Arsenal then finds great settings by comparing the current scene with thousands of professional photos using a Convolutional Deep Neural Network (CDNN).
- Lastly, Arsenal optimizes settings based on 18 different factors, like hyperfocal distance, sensor dynamic range and lens transmission.
Neural Networks Explained
NN are mathematical algorithms that are modeled on the human nervous system. It uses similar processes in the brain for recognizing patterns. The input data (e.g. the image) goes through a process of classification and labeling that determine the settings the device will send to the camera. The output is the exposure and image creation.
The magical transformation takes place when the software recognizes objects and patterns in the image which it compares to its own reference of photos. This uses a technique called CDNN, which analyzes the pixels in the image and based on its training data used in machine learning, allows the software to determine the settings to use for aperture, shutter speed and ISO. This is just a brief explanation in the context of the Arsenal intelligent assistant.
It is all about getting the best out of your existing camera. Don’t expect a cropped sensor to suddenly create far superior images than a full frame sensor. That is not how it works. The most basic way to understand this is that your camera has manufacturer specific features that are provided. What an AI assistant like Arsenal can do is enhance and optimize the best features available to get the best shot possible.
Key features that the device is capable of includes photo stacking for HDR imaging, frame buffers multiple shot selection for sharpest image, better low light shots, as well as finding the best shots in tricky lighting and long exposures. These are always claims that can be at worst debunked or at best proven.
Like with anything new, it has to be tried, tested and true. Arsenal is not a commercial retail product, but is starting out with a limited release that is available from their website. It actually started as a kickstarter project. The aim here is to test the product out in retail and see if this is something that consumers want. Otherwise, it could very well fall under the prosumer market which is more along the lines of professionals and enthusiasts. This may matter more to photographers who want to maximize the best features available from their cameras, but don’t know how to do it. Arsenal, with its AI, can assist and provide the answers on how to improve photography techniques.
For the prosumer market a more premium type of intelligent camera assistant that has more capabilities would be ideal. In this case it is not just about the automation, but the whether the intelligent assistant can make the right decision when it comes to settings and even composition. If it can be left in one spot where it can automatically take photos, like in events, would be a plus.
The benefits from AI doesn’t necessarily have to make humans lazy. What we can learn from it, like what Go players learned from losing to the AI on AlphaGo, are insights on how to become better photographers. There is always plenty of room for everyone to improve their skills. Learning itself is a process, much like how machine learning is helping developers build better products. Arsenal should start the trend of intelligent camera assistants. It would be no surprise if companies like Adobe, Corel and even mainstream camera giant Canon begin building similar devices. If its value is creating better images, then that is what photographers want to get from their camera.
Here is my take though — although there is a lot of automation and enhancement to the image, it doesn’t necessarily mean there is better quality. You can only go so far with AI, but if you compare image quality let us say between a Nikon D5000 using a kit lens and a Canon EOS-5D MarkIV using an EF 85mm f/1.8, the latter will still show that it is a night and day comparison. Quality still depends on the image sensor and lens used on the camera for the type of scene you are capturing.
Note: This review is the author’s opinion based on the publicly available information about the product. The author is not endorsing or being sponsored by the vendor in any way.