Contact a Mobian

We’d love to hear from you!
Drop us a message and we’ll get back to you soon.

Editshare - Enhancing EditShare’s flow media asset management with Automated AI

Editshare - Enhancing EditShare’s flow media asset management with Automated AI

Thought leadership

How computer vision can reduce cart abandonment rates and boost customer revenues

February 16, 2023
April 1, 2022
Clock Icon

From the earliest portraits, to the latest group selfie, photographs have always been at the heart of the human experience. They trigger our emotions, rekindle happy memories and keep us connected with family, friends and loved ones.

But today, finding photographs with emotional significance is harder than ever. The average smartphone owner takes hundreds or even thousands of photographs a year. The plunging cost of storage means there is little incentive to delete unwanted images. Organising them into different events or categories is equally time consuming.

This is the challenge facing internet-to-print businesses looking to grow their revenues. However slick the online experience, customers must still invest time and effort into gathering and organising photographs, even before layout. As a result  of photobooks are abandoned before purchase. Compare this with just 70% across the entire e-commerce sector and it’s easy to see that there are huge opportunities for businesses who can remove friction from the internet-to-print customer journey.  

Bringing emotional intelligence to photobook curation

In recent years some of the big cloud platforms have launched AI services that select photographic highlights and automate photobook layouts. In our opinion their efforts are good but not great. Categorisation is largely generic. Photographs are organised under simple headings such as landscapes, restaurants, sports and so on. Where local legislation allows, facial recognition also plays a role in the filtering process.

But these platforms lack the ‘emotional intelligence’ to interpret the context of similar, but different moments. Riding a bike on holiday in Amsterdam has a completely separate meaning to participating in a cycling road race. Baking a cake for your best friend’s birthday is not the same as eating a slice of your favourite dessert in a cafe.

The algorithms also lack the finesse required to curate a large number of photographs from important life events such as weddings, significant birthdays and holidays.

The perfect marriage of computer vision and photography

At Mobius we have already developed a proof of concept which does just that. In the case of a wedding we showed how it was possible to create a high-quality photo book in just two minutes using a version of our computer vision SDK that has been trained using a large set of wedding photographs.

Once you’ve uploaded your images, the software selects the best shots, based on an aesthetic filter. It then categorises the photos according to common wedding events: the entrance, vows, first kiss, first dance and so on.

Critically, it also uses facial recognition to identify key participants including the bride, groom, bridesmaid and groomsmen. The user can also choose from a thumbnail menu of guests to prioritise pictures containing these people.

Layout is a matter of taste, but again, we trained our algorithms using the best examples of wedding books that we could find. This results in a seamless set of spreads, where each layout matches the mood of the moment.

For instance, the ‘first kiss’ spread features a medium shot of the bride and groom, and then two close ups, during and after the kiss. The guest pages pull together the best pictures of the bride and groom’s family and their friends. At the reception, the software chooses the right balance of images that include highlights from the speeches, catering and dance floor.

Preserving precious memories, long into the future

In the coming months, our computer vision software will play a more active role, correcting lighting and saturation, and cropping images to highlight important details. This is important for special events where you need a good mix of long, medium and close up shots to tell the story of the day.

It will also add a consistent finish to a set of photos from different sources. Where smartphone cameras tend to ‘pop’ the final image adding colour and contrast, specialist cameras capture more nuanced details. You can choose the finish you prefer so that the printed book has a professional look and feel.

Our software also protects the privacy of photobook contributors. The Mobius SDK can be deployed on your laptop or even a smartphone. Photographs from contributors can be stored on the device. There’s no need for a cloud platform that could use your images to target you with advertising or train its own algorithms.

As for our photobook proof of concept, it has already shown its value. As soon as they saw the results, the bride and groom ordered dozens of copies for their family and friends. In the future, computer vision software will be just as precious to internet-to-print businesses looking to reduce abandonment rates and boost long-term customer value.  

Customer details
Written by
Edited by
Illustrated by

Discover more articles

Mobius Labs Unveils Breakthrough Multimodal AI Video Search Empowering Organizations to “Chat” with Their Content Libraries
Turn Back Time with Visual DNA
Thought leadership
Unlocking the True Potential of Content with Visual AI
By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Mobius Labs GmbH is receiving additional funding by the ProFIT program of the Investment Bank of Berlin. The goal of the ProFIT project “Superhuman Vision 2.0 for every application- no code, customizable, on- premise  AI solutions ” is to revolutionize the work with technical images. (f.e.) This project is co-financed by the European Fund for Regional Development (EFRE).

In the ProFIT project, we are exploring models that can recognize various objects and keywords in images and can also detect and segment these objects into specific pixel locations. Furthermore, we are investigating the application of ML algorithms on edge devices, including satellites, mobile phones, and tablets. Additionally, we will explore the combination of multiple modalities, such as audio embedded in videos and the language extracted from that audio.