imgutils.detect.head

Overview:

This module provides functionality for detecting human heads in anime images using YOLOv8 models.

../../_images/head_detect_demo.plot.py.svg

Key Features:

  • Head detection in anime images

  • Support for different model levels (‘s’ for accuracy, ‘n’ for speed)

  • Customizable confidence and IoU thresholds

  • Integration with Hugging Face model repository

The module is based on the ani_face_detection dataset from Roboflow and uses YOLOv8 architecture for object detection.

Example usage and benchmarks are provided in the module overview.

This is an overall benchmark of all the head detect models:

../../_images/head_detect_benchmark.plot.py.svg

detect_heads

imgutils.detect.head.detect_heads(image: str | PathLike | bytes | bytearray | BinaryIO | Image, level: str = 's', model_name: str | None = None, conf_threshold: float = 0.3, iou_threshold: float = 0.7) → List[Tuple[Tuple[int, int, int, int], str, float]][source]

Detect human heads in anime images using YOLOv8 models.

This function applies a pre-trained YOLOv8 model to detect human heads in the given anime image. It supports different model levels and allows customization of detection parameters.

Parameters:

  • image (ImageTyping) – The input image for head detection. Can be a file path, URL, or image data.

  • level (str) – The model level to use. ‘s’ for higher accuracy, ‘n’ for faster speed. Default is ‘s’.

  • model_name (Optional[str]) – Name of the specific YOLO model to use. If not provided, uses default models based on the level.

  • conf_threshold (float) – Confidence threshold for detection results. Only detections with confidence above this value are returned. Default is 0.3.

  • iou_threshold (float) – IoU (Intersection over Union) threshold for non-maximum suppression. Helps in removing overlapping detections. Default is 0.7.

Returns:

A list of detected heads. Each item is a tuple containing: - Bounding box coordinates as (x0, y0, x1, y1) - Class label (always ‘head’ for this function) - Confidence score of the detection

Return type:

List[Tuple[Tuple[int, int, int, int], str, float]]

Raises:

ValueError – If an invalid level is provided or if there’s an issue with the model loading or prediction.

Example:
>>> from imgutils.detect import detect_heads, detection_visualize
>>> image = 'mostima_post.jpg'
>>> result = detect_heads(image)
>>> print(result)
[((29, 441, 204, 584), 'head', 0.7874319553375244),
 ((346, 59, 529, 275), 'head', 0.7510495185852051),
 ((606, 51, 895, 336), 'head', 0.6986488103866577)]

Note

For visualization of results, you can use the imgutils.detect.visual.detection_visualize() function.