3D semantic segmentation by Playment

Accurate 3D point cloud segmentation to train your AI models

Playment's new release of 3D point cloud Segmentation toolkit enables you to generate high-quality training data to build 3D perception models.
Thanks Kevin for hunting us. Hi folks! I'm Sid, cofounder @ Playment. At Playment, we provide high quality training & validation datasets for companies working on computer vision applications. I'm super excited to announce the launch of our latest annotation tool - 3D point cloud Segmentation. This toolkit enables you to generate high-quality training data to build 3D perception models. Computer vision teams send us raw Lidar data and we deliver high quality 3D segmentation training data for your computer vision models. 3D perception modelling has many applications in robotics such as intelligent vehicles, autonomous mapping and navigation. The segmentation process could be helpful for analysing the scene in various aspects such as locating and recognising objects, classification and feature extraction. With focus on accuracy and efficiency in segmenting 3D point cloud we have built unique features like 1. Segmenting using polygons: You can choose a label and draw a polygon enclosing a cluster of points in 3D space which represent that label. All the point within the polygonal pyramid will get that label. This is useful for objects where boundaries are not well defined e.g. trees, vegetation. 2. Segment using cuboid: Draw a cuboid enclosing a cluster of points representing a well defined object. All the points within the cuboid will get that label. 3. Controlling point size: Far away points are usually very sparse and difficult to spot. You can increase the point size to ensure that you have not missed any points.
17 UpvotesShare