Insight Robotics

Collaborative Curation Platform

Project Team: Hubert Thieriot (Product Owner), Justin Lau (Developer), Albert Chan (Developer), Will McCraw (Tech Lead), Hussain Dewani (User Experience)

In order to get teams to work faster, efficiently, and transparently, between different stakeholders the collaborative annotated models they utilize to make sense of, the world must first be interconnected and trained.

Drones are equipped with software including radar and lidar sensors, an array of photo cameras, inertial measurement units (or IMUs), GPS. The drones then capture images requiring users to manually transfer a large amount of data between their drone and computer and that too is only for desktop. Once that data is collected the machine must be taught what to do with it, to make sense of it. A team of manual curators annotates the data by hand, creating models. This is an often costly and inefficient pipeline.

Mid of 2018 I worked with a small team— one product manager, one software architect, and four software engineers—to produce a platform for new, optimized data tools. Our objective was to radically improve the data curation process in order to save time and money, without sacrificing quality in the data being annotated.

As with any design process, we began by conducting research—qualitative surveys of our users, lab studies in our Hong Kong office, partner companies around Asia & North America and competitor analysis. We sought to understand exactly what data our models need to perform successfully, and how curators go about annotating that data.

We found existing tools in the market to be lacking in how they were built and designed. We knew there was an opportunity to improve a front-end experience as well as backend modelling.

To learn more about optimised workflows involving a lot of data, we looked to markets for similar industry tools: photogrammetry and 3D software.

After initial brainstorming and sketching, we began to develop a keen understanding of how our tools would need to work, as well as how we might build a scalable platform for managing and delegating tasks across a team of hundreds of human curators.

Because each task—or annotation scene—is often unique, we needed to enable things like customizing the layout of the tool to allow for multi-view editing of a scene and precise controls of camera angles in a 3D scene. We learned through lab studies that adjusting the layout of the software, saving presets, and giving curators control of their workflow—to meet their diverse visual and cognitive abilities, as well as functional specialties—would enable faster and more efficient work.

Additionally, we intuited that focusing on a keyboard-driven workflow would enable rapid curation of even the most daunting curation tasks, as users wouldn’t need to manually move between keyboard and mouse to accomplish their tasks.

As a result, we were able to develop a high-end photogrammetry software that serve to quickly and efficiently perform map creation from image tiling, multi-level imagery display, and bring all project key players under one collaborative platform.

A/B Timeline View

Project Team: Hubert Thieriot (Product Owner), Justin Lau (Developer), Albert Chan (Developer), Will McCraw (Tech Lead), Hussain Dewani (User Experience)

After 6-months of working on drone data rendering and collaboration software at Insight Robotics, we wanted to accelerate this platform even further by giving our users more clarity and progress of the project with the help of historical data.

We began the project by looking at existing metrics for timeline analysis and data comparison tools. Our initial assumption was that it would be something simple to compare user data between one specific date to another and by simply just storing back-dated data with modern technologies.

We also audited the existing market for web-based photogrammetry platforms and similar experiences across collaboration, productivity in photogrammetry, and urban as well as architectural solutions to get a sense of what experiences were already like.

In mid-March, we launched our initial test to great success. Not only were our users liking the A/B timeline comparison but also indicated it was a much better experience than the outdated tools. We, therefore, integrated A/B Timeline comparison as a key feature - helping the way teams collaborate to make decisions and see their work progress over the period of time.