Labeling video assets

Kili supports all the major video formats and a wide variety of respective codecs. Though we recommend labeling videos imported as native videos, in case of some of the formats, you might have to import your video as separate frames. For more information on the limitations, refer to Adding video assets.

Labeling Video type projects

In your Video projects, you can add these job types.

In object detection jobs added to Video projects, you can use these tools:

• Bounding box
• Point
• Polygon
• Semantic segmentation

📘

Some of these tools have a limited feature set. For more information, refer to the subsequent sections.

Kili's Video interface was designed with focus on readability and ease of use. Kili's video timeline and jobs viewer are designed to clearly show all the annotations that were added and allow easy access to them. You can use the standard set of video controls to:

• Move back and forth between your video frames
• Type a frame number and then jump to a specific frame
• Play the video starting with the frame you’re currently on
• Mute the video

📘

Additionally, you can use Kili's predefined keyboard shortcuts to navigate between frames. For example, you can jump 10 frames forwards and backwards as well as move to the first and last frames of your video. For a full list of options, open your shortcuts menu.

Depending on your ontology and project needs, the recommended labeling techniques might be different (refer to Best practices for labeling). Kili's video interface offers many additional video-specific features and workflows to easily process the labels that were added. For example, after adding an annotation, you can extend it to other frames, adjust it on specific frames, or remove it from other frames.

At any point during the labeling process, you can adjust your environment using the standard UI settings and Kili's dedicated Video controls.

Extending your annotation to other frames

You can easily extend an existing annotation to more frames. You can do that either by dragging the existing annotation across more frames:

or (for example in cases when you need to quickly propagate the existing annotation till the end of your video), use Kili’s Propagate to this frame feature:

Note that you can always shrink your annotation span, too.

Automatically extending your annotation (Smart tracking)

If you want to save yourself some of the manual work, you can use Kili’s Smart tracking feature. Simply select a labeled object, then select Smart tracking from its context menu, and have Kili’s backend automatically label the object on all the remaining frames.

📘

• Currently, Smart tracking only works with the Bounding box tool.
• Depending on what’s easier for you, you can activate Smart tracking from the asset viewer or straight from your video’s timeline.
• Note that smart tracking will only process all the remaining frames in the absence of any key frames. If key frames exist, smart tracking will stop at the next key frame that it finds.

Adjusting your annotation on specific frames

For standard label adjustments, refer to Changing and updating labels.

When you make an adjustment on a specific frame, you create a key frame: a crucial reference point with your object's location and attributes. This information is then used to track the object across the video's remaining frames. Location of any of the objects located in between key frames is inferred by the video labeling engine (automatic interpolation). Key frames are quite significant, as their presence or absence often determines how certain features behave.

📘

• Note that in Kili’s UI, key frames are marked with a square.
• Automatic interpolation works only with bounding boxes and points.

Interpolation works well on linear animations but does poorly on more erratic motions so there will be cases where your moving objects quickly move out of focus. To save you a lot of time and alleviate some of the burden resulting from manual rework, you can use Kili’s Smart tracking feature.

Removing your annotations

If you want to remove your annotation entirely, you can do that directly from the asset viewer, from the job viewer, from the timeline, or by applying a keyboard shortcut. It all depends on your specific use case and preferences.

📘

Note that if you want to remove an item from all the frames, you must use the bin icon accessible from the timeline, to the left of your annotation.

If you want to remove an annotation form specific frames, for example if the Smart tracking mechanism lost track of your object at some point, locate the problematic frames, then right-click them, and select Cut from this frame to remove the annotation from all the consecutive frames.

If our object disappeared only from certain frames, you can use Kili’s Delete selection feature.

If the object re-appears on your video but is labeled as a different item (for example, in cases when you used an ML model to pre-annotate), you can use the Group feature to make sure you have one, instead of multiple objects on your timeline.

If the object corresponds to two separate objects, you can use the Split feature. This will create a new object without losing the annotation work already done.

Adjusting your work environment with dedicated video controls

You can adjust the amount of frames visible in the timeline. If you want a general overview, you can “zoom out”, but if you need to adjust specific annotations on very specific frames, you can “zoom in” to see just a handful of frames in the same timeline and conveniently handle them as separate entities.

By default, your annotations are propagated to the next 100 frames. You can increase and decrease this number, or even switch off the propagation completely, if you prefer to address your frames one by one.

The maximum propagation limit at Kili is 500 frames. If you need more flexibility (for example, you want to propagate your annotation to the end of the video), you can easily extend an existing annotation to more frames.