Glossary

Acquisition frame

One recorded observation in a time series. This workshop labels frames 0 through 29. A frame index is not elapsed time unless the acquisition interval is known.

Additive blending

A display mode that adds light from image layers, helping two channels appear together. Bright overlap can saturate, so inspect each channel separately too.

Array

A rectangular collection of numbers. A microscopy image is represented as an array whose axes may describe time, depth, rows, and columns.

Channel

An image series acquired for a particular signal. The source layer names here are DNA and mitochondria; a layer name alone does not establish the identity of a tracked particle.

Colormap

A rule that maps numerical intensity to displayed color. Changing it does not alter stored intensity values.

Contrast limits

The low and high intensity values mapped to the dark and bright ends of a layer’s display. Values outside the limits may appear clipped.

Detection

A supplied candidate location at one frame. A displayed point does not prove biological identity or tracking correctness.

Downsampling factor

How much the original spatial grid is reduced. Factor 4 keeps roughly every fourth-scale spatial dimension using the preparation method. It reduces workload but loses fine detail. It does not change the 30-frame time axis.

Dtype

The number format stored by an array. uint16 stores non-negative integers from 0 through 65,535.

Environment (.venv)

The folder containing this workshop’s Python and packages. uv sync --locked creates it. It stays inside the workshop folder and does not need manual activation.

Gamma

A nonlinear display adjustment that redistributes apparent brightness between the contrast limits. It changes appearance, not the source values.

Intensity

A numeric image value measured at a voxel. Brightness on screen also depends on contrast, gamma, colormap, opacity, and blending.

Jupyter kernel

The Python process that runs notebook cells. Choose Python (napari workshop) so the notebook uses the workshop environment.

Layer

One image, points table, tracks table, or other object in the napari viewer. Each layer has its own visibility and display controls.

Lazy loading

A strategy that delays reading array chunks until needed. It can reduce RAM use, but it does not shorten the initial download, guarantee instant interaction, or remove GPU limits.

Maximum-intensity projection (MIP)

A 2D image containing the brightest voxel found along a chosen axis, often depth. It shows bright structures together but discards their positions along the projected axis.

Napari

An interactive desktop viewer for multidimensional images and related data. It opens separately from the JupyterLab browser.

Notebook cell

A block of Markdown explanation or Python code. Run code cells from top to bottom with Shift+Enter.

Opacity

How transparent a layer appears. Reducing opacity can reveal another layer beneath it.

Playback fps

Frames shown per second in an exported movie. It controls viewing speed and does not establish acquisition timing.

Point

A marker displayed at a coordinate. In this workshop points visualize precomputed detection-table rows.

Projection

A lower-dimensional summary of an array. Always state which axis and operation were used because a projection loses information.

Scale

The conversion between array steps and physical units. Calibration is not yet confirmed for this dataset, so course coordinates remain in source voxels and no physical scale bar is claimed.

Slice

One plane selected from a volume. Unlike a projection, it does not combine all depth planes.

Source voxel

One element of the original three-dimensional image grid. Tables retain source-voxel coordinates even when the prepared image array is spatially reduced; the course loader aligns them.

Track

An inferred association of positions across frames, identified by id. It is a model result that may contain errors or ambiguities, not a newly measured image channel.

Track tail

The recent portion of a track displayed behind its current point. Tail length describes a window in frame indices, not a particle count.

Voxel

A volume element: the 3D counterpart of a pixel. Its physical size must come from verified calibration.

Zero-based indexing

Counting that starts at zero. Thirty frames therefore use indices 0 through 29.