Dataset for Detection of White Blood Cells of the Juvenile Visayan Warty Pig in Blood Smear Images

This dataset was prepared in part of a study employing YOLOv5 for white blood cell image detection and counting of juvenile Visayan Warty pigs. The dataset is important in supplying material for exploring the application of machine learning for veterinary use. There are a total of 665 annotated 416 x 416 JPG files and their corresponding YOLO Darknet TXT files. The images have been split into training, validation, and testing folders with a 70:15:15 ratio.

---

Folder structure:

1. “test” folder contains two subfolders:
   • "images" containing the 98 annotated images
   • "labels" containing 98 TXT files
2. "train” folder contains two subfolders:
   • "images" containing the 459 annotated images
   • "labels" containing the 459 TXT files
3. “valid” folder contains two subfolders:
   • "images" containing the 98 annotated images
   • "labels" containing 98 TXT files

---

Methodological information:

Date of data collection

June 2021 to November 2021

Sample Collection

Data was collected at Bacolod City, Negros Occidental, Philippines. A smartphone was used to capture images of the peripheral thin blood smears of juvenile Visayan warty pigs viewed under a Olympus® CX23 compound microscope set at 100x Oil Immersion.

Methods for processing the data:

The images were manually annotated using LabelImg.