Glossary

Terms are organized alphabetically

Batch

Number of examples seen per model update step

Cost Function/Loss Function

A lost function is used to calculate the cost, which is the difference between the predicted value and the actual value

Epochs

An epoch is one complete pass through the entire training dataset. More epochs yields better performance at the cost of training time

Excluding mask

Indicates areas of the image set (raw, seg(s)) that will be excluded from training

Image dimension

Dimensionality of your image data, the plugin supports 2D and 3D data

Inference or Prediction

Uses a trained model to make prediction on new data

Learning rate

Scales the magnitude of model weight updates

Loss value

Error between model prediction and ground truth

Merging mask

User-drawn custom shapes, indicateing areas of the base segmentation that will be overwritten by the other segmentation

Model size

Defines the complexity of the model - smaller models train faster, while large models train slower but may learn complex relationships better

Model weight

The parameters learned by the model during training

Patch size

Patch size to split images into during training. Should encompass the structure of interest and all dimensions should be evenly divisble by 4. If 2D, Z can be left blank. Larger patch sizes will take up more memory and have slower training.

Probability map image

Raw model prediction, showing the probability of each voxel belonging to the structure of interest versus background

Raw image

Original microscopy images (.czi, .ome.tiff, .tiff)

Seg 1

Segmentation of the structure of interest (.czi, .ome.tiff, .tiff)

Seg 2 (optional)

(Optional) Complementary segmentation, useful if Seg 1 fails predictably (e.g. a segmentation that works during mitosis to supplement an interphase segmentation)

Segmentation model

A type of ML model that separates the structures of interest from their background in a 2D/3D mircroscopy image

Threshold_otsu

A method for automatically selecting a threshold for binarization. For more information, please see https://scikit-image.org/docs/stable/api/skimage.filters.html#skimage.filters.threshold_otsu

Thresholded binary image

The result from thresholding a probability-map image

Thresholding value

The probability above which a pixel is considered in the foreground

Weights/ Bias

The learnable parameters in a model