Workflow

Using GSense involves following a workflow that utilizes its various functions.

1

Image loading and managing

Load a batch of RGB or hyperspectral images.

2

Spectral Indexing

Compute pseudo-RGB for hyperspectral data.

3

Image Segmentation

Segment images using Segment Anything Model (SAM) to get all masks.

1. Initialize the segmentation model (default or custom).

2. Run segmentation.

4

Annotation (optional)

Refine masks using Napari's editing tools.

5

Binarizer (optional)

Binarize labels

6

Export

Save masks and computed pseudo-RGB images

Watch the below video for a complete workflow:

• Loading 2 hyperspectral images of TIFF file format, 1 jpg RGB image

• Run Segmentation with ViT-h default model

• Binarize generated masks

• Export selected/all layers