Proof of Concept: Wedding Photography Face Recognition

Just over a year ago, I got married to the love of my life. It was the best day of my life.

After our wedding, I was so excited to see all the photos that the photographer had taken to relive the memories, and to share them with our friends and family.

When I finally got the photos, they were beautiful! My friends and family were all excited to see them; however, when I sent them the link where they could view and download them, many of them responded by saying things like "Can you send me just the photos of me somehow?", or "Wow, these are a lot of photos! I want to see just the photos of me and my family, is that possible?"...

And it got me wondering... Why doesn't this happen automatically?

Problem

The problem is quite straightforward: given a set of photos, along with my selfie, return to me the photos which have me in them.

Solution

To build and test such a solution, I use off-the-shelf pre-trained Computer Vision models for identifying a face from one image in a set of other images, in a Python Flask app hosted on Google Cloud Run.

Though this functionality is implemented here in a standalone application, it can just as easily be integrated into a pre-existing application.

Impact

A solution like this has the ability to drastically improve the experience of wedding photography websites' customers. So much so that, considering that this technology has indeed been around for a while, I find myself asking — why haven't such websites already implemented this feature?

Then I realized. There are quite a few stakeholders in the wedding photography business: wedding photography websites, wedding photographers, people getting married, and wedding guests.

The wedding photography website wants to get more wedding photographers to use their software; wedding photographers want to get more people planning a wedding to hire them; and people planning a wedding want their wedding guests to have a great time at their weddings and enjoy the photos thereof!

Added to this, wedding guests are likely to only view the wedding photos a few weeks or even a few months after the wedding, once the photographer is done editing.

With all this, it can be easy for the most upstream stakeholder to forget about the most downstream stakeholder, the wedding guests — even though they may very well be the ones visiting the wedding photography websites the most!

Adding a solution like this to improve wedding guests' user experience can have a trickle-up effect on the whole value-chain:

• If the wedding guests have a good experience, then wedding photographers are more likely to grow their business (e.g. the wedding guests who loved the photos of themselves would hire the same photographer)
• If the wedding photographers grow their business, then they will bring that business to the wedding photography website that they already know and love.
• Everybody wins!