Table 1 Biomechanical and technical problems.

From: Leukocyte Tracking Database, a collection of immune cell tracks from intravital 2-photon microscopy videos

Problem

Description

Effect on cell detection and tracking

Plasticity (Pla)

High variability in cell shape, such as elongation and formation of protrusions

Parts of the same cell not detected or associated to other cells

Contact (Con)

Close proximity of two cells with the same color

Cells merged in a single object. Track interrupted or switched

High background or low signal to noise ratio (BG)

Background or other objects (collagen fibers, auto-fluorescence, cell debris) appear in the same channel of cells with a similar brightness

Inaccurate cell detection, track interruption, tracking of third objects

Fluorescence variation (Flu)

The intensity of fluorescent cells changes during acquisition. Reasons include photo-bleaching and migration in different areas of the tissue

Inaccurate cell detection and track interruption

High migrating velocity (Vel)

Migration velocity greater than the cell size in a time step (absence of overlap)

Track interruption and aliasing if assumptions for interpolation rules for poorly visible cells are not correct. Deformation of cell shapes

Appearance and Disappearance (A/D)

Sudden or progressive appearance/disappearance of a cell, either close to the boundaries of the field of view or in proximity to a blood or lymphatic vessel

Track duration is less or equal than the length of the video. Tracking errors if interpolation rules for poorly visible cells are not correct

Movement of the sample (Mov)

Shifting, drifting or fluctations of the sample due to the movement of the animal or insufficient isolation from breathing, peristalsis and heartbeat

Non-rigid deformation of the tissue, discontinuities in tracks

Microscope instability (Ins)

Noise introduced either by oscillations in the laser power or in the sensitivity of the photo-detectors, resulting in bands or bright spots

Detection of larger or smaller objects. Appearance of the background, disappearance of cells

Large areas (Lar)

Non-uniform brightness

Frequent detection and tracking errors if parameters are not adjusted locally. Increased computational time

Channel specificity (Spe)

Emitted spectrum is captured in more than one channel

Mis-detection and increased contacts with the background

Density (Den)

High number of cells in close proximity

Track switching for tightly interacting cells and mis-detection

  1. Description of the main problems for automatic cell detection and tracking. In brackets it is reported the abbreviation used to refer at each specific problem.