Penguin Image Dataset

This is a dataset of penguin images used to train the YOLOv5 model. The dataset contains images with penguin targets and provides the corresponding annotation information.

Structure of Dataset

In this structure, the dataset is organized as follows:

• data.yaml: The YAML file holds the root path of the entire dataset and the names of the category and the corresponding numerical labels in the dataset.
• coco128/: The root folder of the image data.
  • images/: Contains the image data.
    • train/: Contains the training images.
      • train1.jpg
      • train2.jpg
      • ...
    • test/: Contains the testing images.
      • test1.jpg
      • test2.jpg
      • ...
  • labels/: Contains the label files corresponding to the images.
    • train/: Contains the label files for the training images.
      • train1.txt
      • train2.txt
      • ...
    • test/: Contains the label files for the testing images.
      • test1.txt
      • test2.txt
      • ...

Data Annotation Format

Each image corresponds to an annotation file, using the YOLOv5 annotation format. Each line represents the annotation information for one target object in the format: class_id, x_center, y_center, width, height, as follows:

2 0.41328125 0.81328125 0.4609375 0.3734375

• class_id: the class number of the target object.
• x_center: x-coordinate of the centre of the target object bounding box, scaled relative to the width of the image.
• y_center: y coordinate of the centre of the target object bounding box, scaled with respect to the image height.
• width: the width of the target bounding box, scaled relative to the width of the image.
• height: the height of the target bounding box, scaled relative to the height of the image.

Usage Instructions

1. Download or clone this dataset locally.

git clone https://github.com/Kejia928/local-detection-dataset.git

2. When training a model using the YOLOv5 training script, specify the path to the configuration file for the dataset as data.yaml