Making search easier through the use of faces!
Duration:
0-2 months
Face Recognition is unique feature that users of the DAM opt for when they need to differentiate their images by those located in them. It's importance is not always appreciated until the users see the full benefit. However, like many features it has its flaws.
Updating to React
The system was undergoing a major change; updating Angular to React. With this change it was decided that the whol system would get a face lift. Each feature on the previous version would be enhanced once adopted into the new React system. My job was to understand Face Recognition and to this happen.
Research and Discovery
Face Recognition is not a new technology. Many systems today have it. They work differently depending on where you may find this. Some of the more popular place you might be familiar with are Apple and Google Images (not to mention our top competitors), who subtly add this feature to their images and videos. Users can see what profiles belong to images with people in them, they can assign profiles and even connect profiles to contacts in phones. However, there is so much more work to be done. Many basic functions are not available yet. I wanted to improve this in Face Recognition.
Old Feedback
What did our customers want to see? This was imperative for our product. I went to our data resourses; salesforce, productboard, slack. I collected the reviews/feedback on this feature currently to see what was the general outlook. What that information I next went to our internal team, to see if they could provide me with more insight on what I was seeing.
Much of what was said was the same; simple enhancements that would lift the feature immensely. Some of these included ease of editing profiles, permissions (restricting some users access to certain functions) and removing images from profiles that were incorrectly matched.
Not all these tasks were as simple as they seemed. The backend engineering team were able to give me an estimate on how long all these requests would that. That enabled us to prioritize our work.
Designing Face Recognition
The Interface
The components used throughout the DAM were again utilized here. The use of consistent patterns in our system means training users won't be as challenging since they are familiar with the look and feel already from other features.
The Flow
There is much value in sharing the flows with the engineering team. When designs are finished and handed over to those implementing the changes they can clearly see the intentions behind the CTAs and error flows (should there be any). Clearly labeling flows and adding in correct content, speeds up the building process of the feature.
Mobile and Desktop
Its important for our teams to understand the differences between the desktop and mobile displays. At this stage of our design process we need to establish pattern between mobile and desktop; what users can see and access from both.
General Release
Ease of Use
We wanted our users to be able to easily edit and add their profiles from the image views. In the previous version of the DAM this was not available to users. They needed to go to the main Face Recognition page and edit from there. Users can clearly see how have been identified and unidentified from the page.
Search Enhancements
Searching for identified profiles was never easier. Users can add this criteria to their advanced search in order to find the asset they are looking for. Although it may seem simple this was a feature that took some time as backend work was needed in order to get it running.
Conclusion
Unfortunately some of the features requested were not enabled for the first release of Face Recognition. Due to strict deadlines, some important functions such as permissions and profile editing were not available. As a voice for our users, I try as much as I can to push for change in our features to entice adoption in this amazing feature.