imgalz.models.detector package¶

Submodules¶

imgalz.models.detector.yolov5 module¶

class imgalz.models.detector.yolov5.YOLOv5(model_path, mean=[0, 0, 0], std=[1, 1, 1])[source]¶

Bases: object

YOLOv5 object detection model wrapper using ONNX Runtime.

This class loads a YOLOv5 model in ONNX format and prepares it for inference, including setting preprocessing parameters like mean and std normalization.

model_path¶

Path to the ONNX model file.

Type:: Union[str, Path]

mean¶

Mean values for image normalization.

Type:: List[float]

std¶

Standard deviation values for image normalization.

Type:: List[float]

detect(bgr_img, conf_thres=0.3, iou_thres=0.45, aug=False)[source]¶

imgalz.models.detector.yolov8 module¶

class imgalz.models.detector.yolov8.YOLOv8(model_path, mean=[0, 0, 0], std=[1, 1, 1], nc=80)[source]¶

Bases: YOLOv5

YOLOv8 object detection model wrapper extending YOLOv5.

Inherits basic ONNX loading and preprocessing from YOLOv5, and adds support for specifying the number of classes.

model_path¶

Path to the ONNX model file.

Type:: Union[str, Path]

mean¶

Mean values for normalization. Defaults to [0, 0, 0].

Type:: List[float], optional

std¶

Standard deviation values for normalization. Defaults to [1, 1, 1].

Type:: List[float], optional

nc¶

Number of classes. Defaults to 80.

Type:: int, optional

detect(bgr_img, conf_thres=0.3, iou_thres=0.45, aug=False)[source]¶

imgalz.models.detector.yolov8pose module¶

class imgalz.models.detector.yolov8pose.YOLOv8Pose(model_path, mean=[0, 0, 0], std=[1, 1, 1])[source]¶

Bases: YOLOv8

detect(bgr_img, conf_thres=0.3, iou_thres=0.45)[source]¶

imgalz.models.detector.yolov8seg module¶

class imgalz.models.detector.yolov8seg.YOLOv8Seg(model_path, mean=[0, 0, 0], std=[1, 1, 1], nc=80)[source]¶