By Hans Hartman – Chair – Visual 1st
[Scroll down for And a few more things… industry news highlights]
Photo engagement and monetization go hand in hand – or should go hand in hand, as we heard in an exciting panel at the photokina Business Forum Imaging conference in Cologne last week.
With monetization opportunities abound – from advertising/ sponsorship to selling photo tools/apps or photo print products to offering visual search-enabled shopping solutions, it is easy to ignore the conditions required for monetization to succeed. It’s exactly the challenges of meeting these conditions that are the focus of our panelists’ business models.
Panelists: Philipp Mühlbauer (Picanova), Sofi Shvets (Let’s Enhance), Qian Lin (HP), Ziv Gillat (Athentech)
What is needed to enable monetization of consumers who are taking, sharing or viewing photos in ever-growing numbers? In our panel we discussed 3 important requirements:
1. The photos need to be at the quality required for specific use cases. Ziv Gallat (Athentech, the makers of Perfectly Clear) and Sofia Shvets (Let’s Enhance) represent companies that tackle this challenge from distinctly different angles.
Let’s Enhance uses AI to improve the quality of overly compressed and too low resolution images (such as those shared through texting/messaging apps and various social media sites). This Ukraine-based startup is an offspring of an e-commerce service that suffered from the submission of poor quality images. To tackle this problem Let’s Enhance built neural network-based solutions to automatically remove artifacts and to upscale the photos (2x, 4x, 8x or sometimes even more). Core to its image enhancement solutions is the ability to understand not only what is displayed in the pictures but also what certain items in these pictures (such as skin, hair or clothes) should ideally look like.
Perfectly Clear uses a range of image enhancement algorithms, including those for sceneries and portrait/selfie photos for, for instance, removing wrinkles or enhance skin complexion. Perfectly Clear has been in production for 15 years and is used by many print product providers who save money by automating large scale image enhancement processes. Perfectly Clear applications process 11 billion images a year. Imagine the manpower needed to do this in Photoshop! In the near future, the company will announce a major handset manufacturer who will integrate the Perfectly Clear technology in their phones to improve photos also directly at the point of capture.
The communality between Perfectly Clear and Let’s Enhance? Both companies provide (semi-) automatic solutions to enhance images and sell their solutions as their own B2C apps to end users, as well as through OEM solutions to industry partners.
2. The photos that matter need to be easily discoverable and presentable – ideally done automatically or through the click of a few buttons. While deep learning-based image recognition applications were initially centered around identifying objects within a photo (“find me all the photos with cats”), image recognition today has moved way beyond that use case and provides practical solutions for consumers to more easily and faster create various photo products with the best possible photos presented in the best possible ways. No longer does the consumer need to spend hours to go through their photo collections and to manually lay them out in, for instance, photobooks.
HP. By leveraging deep learning core technologies developed for a variety of business divisions within HP, the group of our panelist Qian Lin has developed modules for creating photobooks that include clustering photos of the same person even when these photos were shot years apart. HP’s modules also analyze the relative importance of certain photos (e.g. based on facial expressions or the direction in which subjects in the photos are looking).
The result is not only automatic selection of photos for the intended photobooks, but also the ability for photobook apps to make smart decisions about which photos go on which page, and which photos are highlighted (placed in the foreground vs. background), while still allowing the user to override these automatic choices.
Not having enough time to create one’s photobook can no longer be an excuse!
3. We need to know the content of what’s in the photos – the more we know who or what is in the photos, the better we can monetize these photos. While this sounds like the credo of any social media company, it’s not a credo that comes natural to photo print product companies. Philipp Mühlbauer of Picanova described his company’s fascinating journey that started as a canvas photo print company to one that explores the frontiers of mass customization through and beyond print.
Picanova stood and stands out by offering a large collection of brands (currently over 40!) through which it sells its wall décor and other print products, but Picanova’s vision goes beyond finite customer segmentation. The company also offers non-photo personalized print products, such as home fashion or textile products, which could include personalized text or designs.
Recently, Picanova launched an automated 3D body scanning service (think of it as a 3D photo booth) at partner retailers through which consumers can receive a free digital file that they can use as an avatar in games or in AR applications. They can also order 3D printed figurines of themselves. Retail partners benefit from this service by attracting more store traffic. Down the road, retailers might also be able to benefit from knowing their customers’ precise body measurements, so that these retailers could reduce returns from online orders or be able to recommend matching clothes. In other words, Picanova and their partners will be able to monetize data derived from the visuals their customers choose to make.
In sum. While visually engaging B2C products or apps can tap directly into a growing variety of monetization solutions, B2B solutions also play an important role in making visual engagement and monetization possible, whether this is by enabling consumers to enhance the quality of their photos so that these become worth using, or by enabling easy curation of the photos that matter, or by using the consumer’s visuals for new and exciting personalized products or e-commerce solutions.
And a few more things…
Allcop. Going WhatsApp: Also at Business Forum Imaging last week: Allcop introduced MagicPostr, a solution developed for German drugstore retailer dm for consumers to order a photo collage poster from inside WhatsApp. Through the WhatsApp messaging bot (if needed supplemented by a human support person) the user can create and order a poster with their favorite photos.
Colourise. Going color: Colourise.sg is a new free site that uses deep learning to smartly colorize your old photos. It was developed in Singapore by the Data Science and Artificial Intelligence Division at GovTech (Government Technology Agency of Singapore). More.
Shutterfly. Going with new leadership: As we described a few weeks ago, Shutterfly’s consumer product division has not been meeting Wall Street’s expectations. So, it is no surprise that the company appointed new leadership for this division: James Hilt, a former executive VP at fashion retailer Express and former managing director of eBooks at Barnes & Noble. More.
Skylum. Going vertical: Skylum announces AirMagic, labeled as the world’s first automated AI-powered software created for drone and aerial photography. AirMagic leverages the company’s existing automatic image enhancement technology. More.
Krome. Going AI: Krome, the company that offers photo enhancement services for consumers and businesses, announces the addition of deep learning-based features to its Krome Business Studiosolution. The deep learning modules leverage Krome’s vast database of images and the work of its crowd-sourced editors to automatically recognize photos and to provides instant design recommendations. More.
Light. Going OEM: Light, you know, the makers of that futuristic smartphone with 16 lenses, announced at MWC the use of its computational photography technology in smartphones from Nokiadevice maker HMD, Chinese handset company Xiamoi and Sony. More.
Nokia/HMD. Going multi-camera: Light’s partner HMD announced the Nokia 9 PureView, a five-camera system developed in close collaboration with Light, Zeiss, Sony, and Qualcomm. Who said that 3 cameras on smartphones were plenty enough? More.
Save the date: Visual 1st: October 3-4, 2019, in San Francisco.
Click here to automatically receive Visual 1st Perspectives in your mailbox