On the radar: Mirrorless AI camera for smartphones

On the radar: Mirrorless AI camera for smartphones

Ever wished for a better camera for the smartphone to take nicer photos? While camera capabilities are limited by the phone’s size, startup Photogram is developing an alternative that brings the benefits of the mirrorless camera to the smartphone.

Its Alice compact, mirrorless camera is sleek and sports just a shutter button, control wheel and cold-shoe adapter. After mounting the smartphone, it connects via 5Ghz wi-fi, letting users share images and stream videos easily.

Alice is built around a Sony IMX294 10.7 megapixel 4/3 sensor optimised for high-quality and full-width 4K video. The sensor is eight times bigger than that of a typical smartphone.

In front of that big sensor is a Micro Four Third lens mount. The compact interchangeable lens system gives users access to more than 100 lens options from Olympus, Panasonic, and specialty lens makers such as Sigma, Tamron and Tokina.

Users will be able to choose from 16mm-equivalent fish-eye lenses for a super-wide angle of view or 800mm-equivalent telephone zoom lenses able to get clear, undistorted pictures of objects far away, with plenty of options in between.

THE SECRET SAUCE

“I was thinking a lot about how data science and machine learning and AI could be applied to create video and imagery. I was using a camera all the time to create video content and becoming increasingly frustrated with operating them,” said Vishal Kumar of Photogram.

The secret sauce is an AI built using NVIDIA GPU-accelerated deep learning to help photographers wring the most out of Alice’s hardware. The AI controls and improves focusing, changes exposure, automatically adjusts white balance, and performs automatic image stabilisation.

Users will eventually make Alice better at whatever photography they do. “Let’s say you’re a wedding photographer, or you like to shoot cats or clothes. We want people to be able to optimise our models and retrain them so their Alice camera can be more optimised for the photography they do,” said Kumar.

Since its listing on Indigogo last September, Alice has raised US$200,000 from more than 250 backers by February. It is due for release in October for around US$760 to early backers.

Photo: Photogram

Countering facial recognition

Source: Sand Lab

If you’re reading this, you are probably using social media as well. One of the biggest worry about social media is the loss of privacy. Every photo that you post of yourself and your loved ones are processed by artificial intelligence (AI) tools that enable facial recognition. In case you’re never wondered why, this is what lets Facebook suggest names to tag when you upload a photo.

Facial recognition works by establishing invisible relationships among the pixels that make up a computer-generated picture of a face. It then compares it with other photos available online to look for matching ones.

This makes tagging photos in Facebook or looking for photos in Google Photos convenient and easy. However, it may also be used by others — organisations and individuals — to do likewise, venturing into the domain of personal privacy.

The good news is that researchers at the University of Chicago have developed an algorithm that counters that. Named “Fawkes“, it alters your photos slightly to deceive the facial recognition software.

Fawkes uses a process called image cloaking to make tiny, pixel-level changes that make it difficult for facial recognition software to connect your photo to other images.

Altered images look the same visually but the subtle change fools facial recognition software, protecting your privacy in the process.

Caveat. I’ve not tried it myself but if you’re keen to check it out, download the free algorithm (for Mac and Windows) at http://sandlab.cs.uchicago.edu/fawkes/#code.

Otter-ly amazing note taking

Otter.JPGTranscribing recordings is one of the most tedious and dreaded tasks. One minute of voice recording can take three or more minutes to transcribe and that’s not talking about recordings that are hard to make out or filled with background noise.

I’ve done this during my air traffic control days when I had to listen to taped transmissions and write down word for word what was communicated between the controller and pilot.

Later on, I had to do this when taking minutes during meetings and as a journalist when playing back interviews.

It is no wonder then that I was intrigued when I read about Otter, a smart note taking software that uses artificial intelligence (AI) to transcribe conversations.

Ten months on and I feel that this app is worthy of a review.

Developed by US-based AISense, the graphics processing unit (GPU)-powered app  records speech and transcribes on the fly when connected to a wireless network. Just looking at the voice-to-text makes me leap with joy.

It’s not 100 percent correct, probably around 70 percent (though the company claims that it’s 90 percent), but that itself saves lots of time. Reasons why it cannot get it fully correct may be because of sound volume or quality, slurring, mispronunciation, or accent.

Otter is free on iOS, Android and the web and comes with 600 minutes of free recording each month. If you need more, you will have to subscribe to a premium plan.

As one who does a fair bit of voice recording, Otter is utterly amazing. It’s a time saver and a dream come true.