update mask2segments and saving results

[vladoossss]

PR for this issue: #9784

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhances segmentation strategies in YOLOv5 by ensuring robust handling of images with no detectable segments.

📊 Key Changes

• Added a conditional check to handle scenarios where no contours are found during segment processing.
• Ensures that a zero-filled array is used as a fallback for images lacking any detectable segments.

🎯 Purpose & Impact

• Purpose: The update aims to prevent errors that occur when the model processes an image without detectable segments by properly handling an empty contour list.
• Impact: This change ensures greater stability for the segmentation functionality in YOLOv5, leading to a more reliable performance for developers and users working with diverse image datasets. 📈

[github-actions]

👋 Hello @vladoossss, thank you for submitting a YOLOv5 🚀 PR! To allow your work to be integrated as seamlessly as possible, we advise you to:

• ✅ Verify your PR is up-to-date with ultralytics/yolov5 master branch. If your PR is behind you can update your code by clicking the 'Update branch' button or by running git pull and git merge master locally.
• ✅ Verify all YOLOv5 Continuous Integration (CI) checks are passing.
• ✅ Reduce changes to the absolute minimum required for your bug fix or feature addition. "It is not daily increase but daily decrease, hack away the unessential. The closer to the source, the less wastage there is." — Bruce Lee

[glenn-jocher]

@vladoossss I'd like to minimize the code changes here. Are you saying that some detections produce no masks?

Is there a way I can reproduce the issue myself? I don't think it makes sense to add integers to a list of segment masks. Can you simply update masks2segments with an if and no else?

def masks2segments(masks, strategy='largest'):
    # Convert masks(n,160,160) into segments(n,xy)
    segments = []
    for x in masks.int().cpu().numpy().astype('uint8'):
        c = cv2.findContours(x, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
        if c:
            if strategy == 'concat':  # concatenate all segments
                c = np.concatenate([x.reshape(-1, 2) for x in c])
            elif strategy == 'largest':  # select largest segment
                c = np.array(c[np.array([len(x) for x in c]).argmax()]).reshape(-1, 2)
            segments.append(c.astype('float32'))
    return segments

[glenn-jocher]

@vladoossss ok took another look. I see we need to keep the output array the same length as the input masks, how about creating an empty nparray when there is no segment of the same 2 columns. Can you try this new function with no other changes anywhere (i.e. use master segments/predict.py etc.)

def masks2segments(masks, strategy='largest'):
    # Convert masks(n,160,160) into segments(n,xy)
    segments = []
    for x in masks.int().cpu().numpy().astype('uint8'):
        c = cv2.findContours(x, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
        if c:
            if strategy == 'concat':  # concatenate all segments
                c = np.concatenate([x.reshape(-1, 2) for x in c])
            elif strategy == 'largest':  # select largest segment
                c = np.array(c[np.array([len(x) for x in c]).argmax()]).reshape(-1, 2)
        else:
            c = np.zeros((0, 2))  # no segments found
        segments.append(c.astype('float32'))
    return segments

[vladoossss]

@vladoossss ok took another look. I see we need to keep the output array the same length as the input masks, how about creating an empty nparray when there is no segment of the same 2 columns. Can you try this new function with no other changes anywhere (i.e. use master segments/predict.py etc.)

def masks2segments(masks, strategy='largest'):
    # Convert masks(n,160,160) into segments(n,xy)
    segments = []
    for x in masks.int().cpu().numpy().astype('uint8'):
        c = cv2.findContours(x, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
        if c:
            if strategy == 'concat':  # concatenate all segments
                c = np.concatenate([x.reshape(-1, 2) for x in c])
            elif strategy == 'largest':  # select largest segment
                c = np.array(c[np.array([len(x) for x in c]).argmax()]).reshape(-1, 2)
        else:
            c = np.zeros((0, 2))  # no segments found
        segments.append(c.astype('float32'))
    return segments

Yes, modifying just this one function also works fine!

[glenn-jocher]

@vladoossss great! PR is merged. Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐