Auto Track

Usage  |  Controls

The Auto Track node can be used to automatically generate trackers to use for camera or object solving, and has a single input and output. It attempts to pick suitable points in each frame and match them together to form a set of connected feature tracks.

The Auto Track node is designed to work best with features picked from natural images, such as marks on a wall, or stone or fabric texture for example. For scenes containing more synthetic features such as large tracking markers on a green-screen set, the User Track node can be used for more manual control of exactly where the trackers are generated.

Tracking groups

All trackers generated by a single Auto Track node are placed into a tracking group. By default, this group is called Camera and corresponds to tracking the overall camera motion. When tracking an object, the group should be changed to something representative of the object being tracked (for example, "Car" or "Head").

Usage

The Auto Track node works by first identifying a large number of candidate points in each frame, and matching those points from frame-to-frame over the entire clip. Once the candidate points have been matched, they are filtered down into a smaller set of trackers which are then passed down-stream.

The trackers are chosen so that they best represent the overall contents in the shot, exist for a large number of frames, and are well distributed over the image area. The number of trackers that are generated can be adjusted interactively without the need to re-generate all the candidate points each time, greatly speeding up the process of selecting suitable parameters for a shot.

Clicking the Start Tracking button will start the tracking process.

After trackers have been generated, their image motion and tracking length are presented in the Motion and Coverage graphs respectively. Further information about these tools is available below.

Trackers can also be selected in the Cinema window by clicking with the left mouse button. Multiple trackers can be selected by holding the Ctrl key whilst clicking.

Candidate Points

Candidate points in each frame are selected by looking at the image data and identifying areas that correspond to strong features such as corners or image texture. The number of candidate points that are then selected for matching in each frame is controlled by two main parameters in the Parameters panel:

- The Feature Scale value determines the overall scale of features that examined. When this is set to Small, small corners and image texture regions will be used. Large means larger scale features will be selected. Looking for features at a larger scale can be beneficial in situations where image noise or other factors have a detrimental effect on the selection of candidate points, but may reduce the overall accuracy of the solve as large points can be more difficult to locate precisely in other frames.

- The Feature Confidence value controls the number of points in the image that will be considered for tracking. A larger value means only the features with a high confidence will be used. A lower values means features with a lower confidence will be picke dand tracked, although these features may not track as accuractely as those with a high confidence.

- The Motion Prediction parameter defines the technique used to predict how points are moving from frame-to-frame. The default value of Using Image will use the overall image motion as a predictor.

- The Search Area parameters define the size of the search window used when trying to locate where a point moves from one frame to the next. This defaults to a fixed percentage of the image size, but can be reduced or increased if necessary in situations where there is very little or very large amounts of image motion. The search area should be set large enough to ensure features can be matched from frame-to-frame.

Previewing Trackers

To see how the tracking parameters will affect the point that are picked as candidate points, click the Preview Tracking button. This will display dots in the Cinema window that are coloured according to their status:

• Points that have by picked and tracked successfully are shown in yellow
• Points that have by picked for tracking but could not be tracked are show in red
• Points that have not been picked for tracking because their confidence was below the Feature Confidence threshold are shown in purple

For points that are tracked successfully, their motion vector is also drawn as a yellow line. You can also preview the Search Area size for trackers by enabling the Preview Search Windows option and hovering the mouse over a tracking point. Use this to judge whether the search area is too large or too small for your clip.

You can use the preview tool to help adjust the Feature Confidence parameter to ensure enough are picked over your entire image which will help to produce a more accurate camera solve.

Masks

Masks can be generated to restrict candidate point selection to a specific part of the image. This is useful when the clip contains objects moving independently from the camera, and points are being generated for either camera or object tracking.

To mask candidate points, connect either a Roto or Keyer nodeto the Auto Track node as described in the Workpage section of this documentation, Or load an Image-based mask into the upstream Clip Input node.

Disabling trackers

Individual trackers can be disabled where required (for example, if they have been tracked incorrectly) by clicking the Disable button and then clicking on tracking points in the Cinema window with the left mouse button. Clicking and dragging with the left mouse button will draw a rectangular selection marquee. Holding the Alt/Option key will also allow a lasso to be drawn. All trackers inside the marquee will be disabled.

Disabled trackers can be re-enabled at a later date if required by clicking the Enable button. This will show all disabled trackers in the Cinema window in orange alongside the other trackers that are still enabled. Clicking with the left mouse button (or using the marquee tools again) will enable these trackers. Frames where trackers are disabled will also be highlighted in the scrub bar.

Editing trackers

Trackers can be made available downstream for further adjustments by attaching a User Track node beneath the Auto Track node, and clicking the Fetch button to fetch individual trackers for editing. See the User Track node documentation for further details.

The Coverage Graph

The coverage graph displays information about the number of trackers that have been generated for each frame of the clip. The graph displays a colour-coded count of the number of trackers in each frame:

The graph is colour-coded according to the length of each tracker. Trackers that exist for 5 frames or less are coloured red, and points that exist for more than 50 frames are coloured green, with graduations of orange and yellow between.

Below the graph is a numerical breakdown of the overall length of each tracking point, including the average track length, and a percentage breakdown of the entire set of tracking points, colour-coded using the same scheme as the graph.

This graph can be used to quickly identify frames where further attention may be required because fewer features could be tracked, or frames where the number of frames each tracker exists for is small. In these cases, adding a User Track node downstream and manually tracking additional points may help.

Mouse controls

Hovering the mouse over the coverage graph will display a tooltip showing the statistics for the current frame.

To pan the graph horizontally, click and drag with the Right mouse button. To zoom, click and drag with the Middle mouse button (these buttons can be changed in the User Preferences window). Alternatively, the mouse wheel can be used to zoom as well if desired.

The Motion Graph

The motion graph displays plots for the horizontal and vertical motion of each tracker in each frame. The graph can display either tracker velocity or acceleration by selecting the appropriate option from the menu below the graph.

The velocity and acceleration graphs can be used to help quickly identify trackers that are moving differently compared to the rest of the image. For example, this may be because the trackers exists on an object that is moving independently from the camera, or it may be because the tracker is positioned on a false corner, or on another point that cannot be tracked accurately. These situations will be shown as jumps and sharp changes in the velocity or acceleration of the tracker.

Horizontal motion is shown in the upper graph in red, and vertical motion is shown in the lower graph in green. The motion of any selected trackers is shown in blue.

Below the graph is a menu that can be used to switch between displaying tracker velocity or acceleration. At the bottom-right is a button that will fit the graph viewport to display the entire clip (or the range of frames occupied by any selected trackers).

Mouse controls

To pan the graph horizontally, click and drag with the Right mouse button. To zoom, click and drag with the Middle mouse button (these buttons can be changed in the User Preferences window). The graphs can be panned or zoomed both horizontally and vertically by dragging the mouse in the appropriate direction.

Alternatively, the mouse wheel can be used to zoom as well if desired. Scrolling the wheel will zoom the graph horizontally. To zoom vertically, hold the Alt/Option key whilst scrolling the wheel.

The Disable button that is used to disable trackers in the Cinema window can also be used to disable trackers in the motion graph. Clicking the Disable button and then clicking on a tracker (or clicking and dragging to use the marquee tool) will disable that tracker and remove it from the graph and the Cinema window. To re-enable the tracker, click the Enable button and click on the tracker in the Cinema window. Trackers cannot be permanently deleted, only disabled.

Controls

Tracking Group

Create a new tracking group.

Current group: Specify the group that new tracking points will be associated with. Groups can be renamed by typing into the menu box.

Delete the current tracking group.

From: Set the first frame to track to the current frame. The corresponding frame number is displayed in the edit box. Reset with button.

To: The last frame to track to the current frame number. The corresponding frame number is displayed in the edit box. Reset with button.

Trackers

Target number: Target number of trackers to generate for each frame.

Min length: The minimum number of frames that each tracker must be tracked. Increasing this number will remove short trackers and generate more longer trackers if possible.

Disable: Click this button and then click in either the Cinema window or Motion graph to disable a tracker. Multiple trackers can be disabled by clicking and dragging to draw a selection marquee.

Enable: Click this button to display disabled trackers in the Cinema window in orange. Click on disabled trackers to re-enable them. Multiple trackers can be enabled by clicking and dragging to draw a selection marquee.

Parameters

Store current parameters as a new preset.

Current preset: Current preset for the tracking parameters. The current preset name can be changed by typing into the menu box.

Delete the current preset.

Channels: Image and mask channels to use for point selection and tracking.

Feature scale: Scale of features expected in the image. When set to Small (the default value), small-scale corners and image features will be picked and tracked. This will generally provide the best accuracy, but if only larger features exist in the shot (for example, due to focus issues) the using a Large feature scale can sometime help to pick them out more accurately.

Feature confidence %: This value controls the number of points in the image that will be considered for tracking. The default value is 50%. Decreasing this value will mean more points are selected for tracking, but these points may not track as well as others that have a higher confidence rating.

Search area: The width and height of the search area (in pixels) used when attempting to track points in one frame to another.

Feature template: This parameter specifies whether the feature template that is used for tracking should remain fixed (the default), or should change from frame-to-frame as the clip is tracked (Adaptive). A fixed template will generally produce a more accurate set of tracking points, and an adaptive template will often produce longer tracks. However, an adaptive template can also introduce more errors into your trackers that may need further editing.

Refine: This option controls whether an additional refinement process should be performed for each tracker that estimates how the feature template warps from frame-to-frame. Enabling this option will generally produce a more accurate set of trackers, but comes at the cost of a small increase in the time required to track each frame.

Motion prediction: The method used to predict where one feature is moving to in the next frame. The default value is Image Motion, indicating that the average motion observed over the 2D image should be used. Alternatively, setting this to From Metadata will use metadata describing approximate camera motion if that is available for the clip. Setting this to None means that no prediction is used, which can often speed up the tracking process but may mean some points are not tracked as accurately.

Reset: Reset the tracking parameters back to their default values.

Edit ROI: Adjust the region of interest (ROI) for matching in the Cinema window. Click and drag with the left mouse button to adjust edges of the ROI. Points outside the ROI will be rejected during the matching process.

Skip Frame: Skip the current frame during feature matching. Once skipped, the button label will change to Un-Skip Frame to include the frame during the matching process. Skipping frames can be useful when one frame is missing or corrupted due to some sort of image degradation.

Start Tracking: Start the automatic tracking process, first tracking candidate features throughout the clip and then automatically filtering them to generate a set of trackers. Holding the Shift key whilst clicking this button will run the tracking process in the background.

Display

Centre View: Translate the image displayed in the Cinema window so that the selected tracker is positioned in the centre of the window. This can be useful when checking the motion of one particular tracker to see how well it manages to track an image feature over multiple frames.

Preview Tracking: When enabled, candidate point locations will be displayed in the Cinema window. This can be used to quickly assess how different parameters affect the features before tracking. Hovering the mouse near a preview point will display a numerical readout of the tracking motion to the next frame.

Preview search windows: When enabled, hovering the mouse over a point will also display the Search Window that was used to track the feature into the next frame.

Preview failures: When enabled, candidate features that were not tracked are also displayed in red.

Show Tracking in Cinema: When enabled, all the Cinema will be updated whilst tracking is being performed. Disabling this option may increase overall tracking speed.

Show Paths: When enabled, all tracker paths will be displayed in the Cinema window.

Show Information: When enabled, pixel coordinate information for each tracker will be displayed in the Cinema window.

Default Keyboard Shortcuts

Keyboard shortcuts can be customised in the Preferences.

Start Tracking	Shift+A
Show Paths	Shift+P
Show Information	Shift+I
Show Tracking in Cinema	Shift+U
Skip/Un-Skip Frame	Ctrl+S
Preview Tracking	Ctrl+P
Preview Search Windows	Ctrl+T
Preview Failures	Ctrl+F
Centre View	Shift+C