Zeyu Xu

Let's enjoy AI and remote sensing!

How to use the “Wildlife Detection Model”

How to Use the “Wildlife Detection Model” https://ai.zeyuxu.com/#detection The Basic Detection Model is a streamlined wildlife detection system optimized for standard-sized images (512×512 pixels). It excels at rapid identification of five distinct wildlife categories: White-eared Kob, Mongalla Gazelle, Tiang, Reedbuck, and other wildlife species. Each detection is visually represented with color-coded bounding boxes – red for White-eared Kob, green for Mongalla Gazelle, cyan for Tiang, pink for Reedbuck, […]

How to Use the “Wildlife Detection Model”

https://ai.zeyuxu.com/#detection

  1. You can directly drag and drop images into the interface.
  2. Click “Start Processing”. At this point, the threshold on the right represents the minimum confidence threshold.
  3. You can view the results based on the confidence threshold. Only detections with confidence scores higher than the threshold at the time of clicking “Start Processing” will be displayed.
  4. You can download the results in YOLO format.
  5. You can remove an image and drag it back in again

The Basic Detection Model is a streamlined wildlife detection system optimized for standard-sized images (512×512 pixels). It excels at rapid identification of five distinct wildlife categories: White-eared Kob, Mongalla Gazelle, Tiang, Reedbuck, and other wildlife species. Each detection is visually represented with color-coded bounding boxes – red for White-eared Kob, green for Mongalla Gazelle, cyan for Tiang, pink for Reedbuck, and blue for others. The interface features an intuitive confidence threshold slider that users can adjust in real-time to filter detection results based on confidence scores. The system provides immediate visual feedback and detailed statistics about detected objects, including precise confidence percentages for each identification. All detection results are automatically saved in YOLO format text files, making them compatible with various wildlife research and analysis tools. The system also includes a convenient drag-and-drop interface and supports multiple image formats including JPG and PNG.