Functions by annotation product
Updated at July 9th, 2024
Here at Sama, we offer 4 tools that form the Sama Annotate platform for a comprehensive experience:
- Image annotation: Sama image annotation allows you to provide image assets to be labeled through either shape or semantic segmentation tools.
- Video annotation: Sama video annotation experts precisely track temporal and spatial features of your target objects and improve your machine learning model’s performance.
- 3D Point Cloud annotation: Sama 3D point cloud annotation combines the capabilities of fully designed cuboids, lines, and polygons to enable comprehensive spatial and temporal ground truth annotations.
- Sensor Fusion annotation: Sama sensor fusion annotation allows the simultaneous annotation of multiple distinct sensor data points by combining 3D Point Cloud and Video annotation.
The table below maps out what each annotation platform offers and how you can leverage Sama for your data.
| Annotation Platform | Bounding Boxes | Keypoints | Polygons | Lines and arrows | Cuboids | Semantic Segmentation | Temporal tracking | Spatial tracking |
|---|---|---|---|---|---|---|---|---|
| Image |  |
 |
|
|
 |
 Pixel or Polygon |
|
|
| Video | |
|
|
|
|
 Polygon |
|
|
| 3D Point Cloud | |
|
|
|
|
|
|
|
| 3D Sensor Fusion | |
 Works in Video, not in 3D Point Cloud |
|
|
|
 Works in Video (Polygon), not in 3D Point Cloud |
|
|
Supported formats
When you're setting up a workspace output, you will usually need to map it to a URL input to bring in assets. The table below details the supported file formats and maximum file sizes for each workspace output, and provides examples.
| Workspace output | Supported URL inputs | Max. file size | Sample file | Notes |
|---|---|---|---|---|
| Image Annotation | .jpeg, .jpg, .png | ≤ 2 MB hosted on our S3 bucket or a client-hosted public URL, or ≤ 10 GB hosted on local asset servers | .jpg | |
| Video | Video files (.avi, .mp4, .mpg, .mov) or Image frames in subfolders or a .zip file with the following sequence: 0.jpg, 1.jpg, 2.jpg, etc. |
≤ 25 GB in one workspace across all outputs |
.zip with image frames .mp4 |
|
| Point Cloud | .zip file with the following sequence structure: frame001.pcd, frame002.pcd, frame003.pcd, etc. The full list of supported files is here. Folders are also supported instead of .zips as inputs. However, the folders (and the files within) must be hosted in the Sama S3 bucket. |
.zip with .pcd files | Each data point should have (x,y,z) coordinates. Point intensity values are also optionally supported (normalized to a 0-1 scale), encoded in an intensity dimension. The positive z direction must represent “up” in the physical world. Intensity values can also be normalized with a pre-processing setting. |