Keymakr partners with Segments.ai to boost 3D point cloud annotation capabilities
Keymakr announces a strategic partnership with Segments.ai, a leading labeling platform designed for 3D point clouds and multi-sensor formats. This collaboration aims to streamline labeling workflows, enhance accuracy, and reduce the time required to prepare large datasets, a critical factor for autonomous driving and other industries relying on 3D data.
Why 3D labeling matters
3D point cloud annotation is a cornerstone technology for computer vision applications, including autonomous vehicles and ADAS, off-road, delivery, maritime robotics, and medical imaging. Unlike 2D images, which offer only a flat representation, 3D point clouds capture the spatial position of objects, enabling a full reconstruction of the environment.
Yet, working with 3D data is challenging. LiDARs and radars generate massive datasets that can overwhelm even the most powerful systems. Artifacts, such as reflections from glass or water, create “phantom” points, introducing objects that do not exist in reality. As a result, model performance is highly dependent on annotation quality: even minor errors or omissions can compromise AI training.
Partnership in action
To tackle these challenges, Keymakr partnered with Segments.ai, a leading labeling platform built for multi-sensor annotation tasks. Segments.ai enables seamless integration of data from different sensors and ensures consistent and accurate annotations across time and modalities.
This partnership allows Keymakr to integrate tools for robust 3D data handling into their workflows and optimize key stages of annotation.
Core features driving this improvement include:
- Support for diverse 3D formats, including uncommon formats frequently encountered in real-world projects
- Vector-based annotation tools significantly reducing the time required for 3D tasks
- Powerful interpolation tools for moving objects with cuboids
- Collaborative workflows to simplify team management and streamline the process
- Technical support for unconventional materials, with Segments.ai engineers assisting in format adaptation and data import
A win-win collaboration
“We carefully selected a partner for 3D data annotation. Segments.ai’s combination of tools gave us the most efficient results, reducing annotation time by about 20% compared to competitors. Their platform continues to provide the best tools and conditions for our team.”
Michael Abramov, CEO at Keymakr
Keymakr annotators now work directly in the Segments.ai environment. Beyond standard tools, the team implemented a scalable workflow allowing multiple specialists to work on the same file simultaneously. Tasks that previously took hundreds of hours of sequential work are now distributed across a team, producing results tens of times faster.
This flexibility allows Keymakr to meet client needs efficiently on large-scale projects.
For Segments.ai, collaboration with Keymakr offers an additional opportunity to test tools on substantial datasets in autonomous vehicles and other industries. “We are always exploring creative ways to leverage our partners’ tools because our customers often request non-standard and challenging tasks,” says Zoya Boyko, Keymakr Project Manager. Working with unconventional formats and complex cases helps Segments refine its platform, making it more versatile.
"Keymakr’s team makes full use of our platform to optimize speed and quality across complex projects. Their proactive communication and expert feedback help us continuously improve the platform. It’s always great to see the results they’re achieving for some of the customers we support together."
Otto Debals, CEO at Segments.ai
Future prospects
Looking ahead, the future of 3D labeling is tied to several promising directions. As multimodal AI models become more widespread, annotation will expand beyond visual data to integrate LiDAR, radar, audio, and even haptic inputs, requiring richer and more contextual labels.
High-precision 3D annotation still relies on the expertise of human annotators, but their work is now amplified by advanced ML-powered tools. Automation helps accelerate routine steps, such as handling cuboids, while expert verification ensures that the final datasets meet the highest quality standards. This synergy allows teams to deliver results that are both faster and more reliable.
3D segmentation, in particular, highlights the need for this balance. Complex scenarios, like distinguishing a person from a bicycle they are riding, require not just careful manual judgment but also powerful tool support to streamline the process. Together, these capabilities reduce effort without compromising accuracy.
While 100% automation is not yet possible, combining Segments.ai’s advanced platform with Keymakr’s proven expertise creates a best-of-both-worlds solution: smart tools that accelerate work, backed by human insight that guarantees precision. This partnership enables the creation of scalable, accurate, and future-ready 3D datasets.
