Classified bounding boxes of 2D fluorescent microglia images

This data set contains 126 pairs (image + set of classified bounding boxes) of annotated images representing 2D fluorescent microglia. These images are squares of 512x512 pixels, coming from acquisitions in 2048x2048 pixels. 

This data was used to train a YOLOv5 model.

inputs.zip: Contains PNG images. They were normalized in the range 0 to 255 from the min and max found in the original image.

labels.zip: Contains TXT files representing classified bounding boxes, following the format expected by the YOLOv5 model. It means that each line represents a tuple of length 5: classIndex, xMin, xMax, yMin, yMax. Each of these components is expressed in the percentage of the width/height of the image.

classes.txt: Litteral name of the classes represented by indices in the bounding boxes.

Microscope
ZEISS Axio Imager 2

Objective
20x (Air)

Acquisition mode
Widefield

Fluorophore
eGFP

Pixel size
0.325 µm

Dimensions
2D, single channel