Spot detectors are TrackMate modules that detect objects in an image. They are detection or segmentation algorithms that produce objects that will be later filtered then tracked with other TrackMate modules.
In TrackMate, there is two kind of detectors:

Spot detectors. They can detect objects, that is: they output spots characterized only by their X, Y, Z position and a radius. These spots have no shape information. The detection algorithms are typically fast and robust and work in 2D and 3D indifferently.

Segmenters. These detectors are a novelty introduced with TrackMate v7. They can segment objects, that is: they output the objects shape which can be used later to get morphological descriptors and accurate intensity measurements in objects. As of today, TrackMate is limited to segmenting objects for 2D images. Some segmenters work on 3D images, but output objects as shapeless spots.
Spot detectors.
 Laplacian of Gaussian (LoG) detector.
 Difference of Gaussian (DoG) detector.
Segmenters.
The new detectors introduced in the version 7 of TrackMate and that can segment objects in 2D are introduced here.
Individual documentation pages:
 Mask detector: trackmatemaskdetector
 Thresholding detector: trackmatethresholdingdetector
 Label image detector: trackmatelabelimagedetector
 TrackMateIlastik: trackmateilastik.
 TrackMateMorphoLibJ: trackmatemorpholibj
 TrackMateStarDist: trackmatestardist
 TrackMateCellpose: trackmatecellpose
 TrackMateWeka: trackmateweka
Spot features generated by the spot detectors
Features in TrackMate are numerical values defined for one of the TrackMate objects. All detectors must at least provide the following common spot features:
X
,Y
, andZ
: the spot coordinates in space. Coordinates are expressed in physical units (μm, …).R
the spot radius, also in physical units. The current detectors only set this radius value to be the one specified by the user. More advanced detectors  yet to be implemented  could retrieve each spot radius from the raw image.Quality
: The implementation varies greatly from detector to detector, but this value reflects the quality of automated detection. It must be a positive real number, large values indicating good confidence in detection result for the target spot. This sole feature is then used in the initial filtering step, where spots with a quality lower that a specified threshold are purely and simply discarded.
The two other time features  T
and FRAME
number  are set by TrackMate itself when performing detection on all the timepoints of the target raw data. T
is expressed in physical units, and the FRAME
number  starting from 0  is simply the frame the spot was found in.
Object morphology analysis
One of the main goal of generating the object shape is to measure their shape. Detectors that return object contours trigger automatically the computation of morphological features in TrackMate. When relevant, these features use the physical calibration and units of the source image. This means for instance that if your image is calibrated with a pixel size in µm, the area of objects will be expressed in µm². These morphological features are:
Area
Simply the area of the objects in spatial unit squared.
Perimeter
The length of the contour in spatial unit.
Circularity
The circularity is a measure of how close to a circle the shape of an object is. It has a value of 1 for circles and is getting close to 0 for very elongated objects. It is computed for 2D objects as
\[\frac{4 \times \pi \times \text{area}}{\text{perimeter}^2}\]Solidity
The solidity ranges from 0 to 1 and reports how smooth and convex is the object contour. A object with a dented contour, with many cavities will have a low solidity, close to 0. A perfectly convex object will have a value of 1. To compute it we first determine the convex hull of the object. Intuitively, this is the contour we would get if we would wrap a rubber band around the object. It would stretch around the object contour, and would not extend inside the cavities of the object. The area of this convex object is therefore always larger than the area of the initial objet. Then the solidity is computed as:
\[\text{solidity} = \frac{\text{area}}{\text{convex area}}\]Ellipse fit
Several 2D morphological features are best obtained by fitting on ellipse on the object contour, and reporting the ellipse parameters. In TrackMate, we first fit an ellipse to the contour using a direct fit following the Chernov method, computed using the MoorePenrose pseudo inverse (by Kim van der Linde) for speed and robustness. TrackMate then reports the resulting ellipse parameters:
Ellipse center x0
and Ellipse center y0
This the ellipse center position, with respect to the object center position. You can get the object center position by using
double x = spot.getDoublePosition( 0 );
double y = spot.getDoublePosition( 1 );
to which you need to add the ellipse center values to have the absolute position of the ellipse center.
Ellipse long axis
and Ellipse short axis
The length of the long and short axix of the ellipse, in physical units.
Ellipse angle
The angle of the ellipse long axis with the Xaxis of the image, in radians. Careful, in images the Y axis runs from top to bottow, so the positive angle direction is inverted compared to classical plots (positive angles are counterclockwise) on the image.
Ellipse aspect ratio
The ellipse aspect ratio is the ratio of the major axis to the minor axis:
\[\text{ellipse AR} = \frac{\text{major axis}}{\text{minor axis}}\]It ranges from 1 for ellipses that resembles circles, and gets larger for elongated ellipses. A perfect line as a positive infinite aspect ratio.