Is it possible to change the order of detection from top to bottom?

[Flavien123]

I have a vertical image and after the save_crop inference, the images are in order from bottom to top. Is it possible to change it like this? I use CLI commands with custom trained YOLOv11 model

[UltralyticsAssistant]

👋 Hello @Flavien123, thank you for your interest in Ultralytics 🚀! We recommend a visit to the Docs for new users where you can find many Python and CLI usage examples, and where many of the most common questions may already be answered.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to help us better understand and debug your issue.

If this is a ❓ Question regarding custom usage, such as how to modify detection order for your specific vertical image, please provide additional details about your setup. This might include your CLI command, model training configuration, and any adjustments you've already made. If possible, include relevant script snippets or settings to help clarify your usage.

Join the Ultralytics community where it suits you best. For real-time chat, head to Discord 🎧. Prefer in-depth discussions? Check out Discourse. Or dive into threads on our Subreddit to share knowledge with the community.

Upgrade

Ensure you're running the latest ultralytics package to leverage all the latest bug fixes and features in a Python>=3.8 environment with PyTorch>=1.8. You can upgrade using:

pip install -U ultralytics

Environments

YOLO may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all Ultralytics CI tests are currently passing. CI tests verify correct operation of all YOLO Modes and Tasks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

This is an automated response to help guide you. Don't worry though, an Ultralytics engineer will follow up to assist you further as soon as possible 😊.

[glenn-jocher]

@Flavien123 currently, the order of detections in save_crop is determined by the bounding box positions processed during inference, and there is no built-in option to reorder them directly in the CLI. However, you can customize the order programmatically by post-processing the results in Python. Let me know if you'd like an example of how to do this!

[Flavien123]

Yes, is it possible to have a Python code example where the order will go from right to left and down with save_crop?

[glenn-jocher]

@Flavien123 certainly! Here's a Python example to reorder the crops from right to left and top to bottom based on their bounding box coordinates:

from pathlib import Path
from ultralytics import YOLO

model = YOLO("path/to/your/custom_model.pt")
results = model("path/to/your/image.jpg")

# Define save directory
save_dir = Path("path/to/save/crops")
save_dir.mkdir(parents=True, exist_ok=True)

# Extract boxes and sort (right to left, top to bottom)
detections = results[0].boxes
sorted_boxes = sorted(detections, key=lambda x: (-x.xyxy[2], x.xyxy[1]))  # Sort by x2 descending, then y1 ascending

# Save crops in the specified order
for i, box in enumerate(sorted_boxes):
    save_path = save_dir / f"crop_{i + 1}.jpg"
    box.save_crop(file=save_path, img=results[0].orig_img)

print(f"Crops saved in {save_dir}")

This sorts detections by the rightmost x-coordinate (x2, descending) and then by the top y-coordinate (y1, ascending). Let me know if you need further clarification!

[Flavien123]

Thanks, its work!

[glenn-jocher]

You're welcome! Glad it worked for you—credit goes to the YOLO community for making it possible. Let us know if you have more questions!