Function annotate

agents/ui_agent/util/utils.py:326–354

This function annotates an image with bounding boxes and labels. Parameters: image_source (np.ndarray): The source image to be annotated. boxes (torch.Tensor): A tensor containing bounding box coordinates. in cxcywh format, pixel scale logits (torch.Tensor): A tensor contai

(image_source: np.ndarray, boxes: torch.Tensor, logits: torch.Tensor, phrases: List[str], text_scale: float, 
             text_padding=5, text_thickness=2, thickness=3)

def annotate(image_source: np.ndarray, boxes: torch.Tensor, logits: torch.Tensor, phrases: List[str], text_scale: float, 
             text_padding=5, text_thickness=2, thickness=3) -> np.ndarray:
    """    
    This function annotates an image with bounding boxes and labels.

    Parameters:
    image_source (np.ndarray): The source image to be annotated.
    boxes (torch.Tensor): A tensor containing bounding box coordinates. in cxcywh format, pixel scale
    logits (torch.Tensor): A tensor containing confidence scores for each bounding box.
    phrases (List[str]): A list of labels for each bounding box.
    text_scale (float): The scale of the text to be displayed. 0.8 for mobile/web, 0.3 for desktop # 0.4 for mind2web

    Returns:
    np.ndarray: The annotated image.
    """
    h, w, _ = image_source.shape
    boxes = boxes * torch.Tensor([w, h, w, h])
    xyxy = box_convert(boxes=boxes, in_fmt="cxcywh", out_fmt="xyxy").numpy()
    xywh = box_convert(boxes=boxes, in_fmt="cxcywh", out_fmt="xywh").numpy()
    detections = sv.Detections(xyxy=xyxy)

    labels = [f"{phrase}" for phrase in range(boxes.shape[0])]

    box_annotator = BoxAnnotator(text_scale=text_scale, text_padding=text_padding,text_thickness=text_thickness,thickness=thickness) # 0.8 for mobile/web, 0.3 for desktop # 0.4 for mind2web
    annotated_frame = image_source.copy()
    annotated_frame = box_annotator.annotate(scene=annotated_frame, detections=detections, labels=labels, image_size=(w,h))

    label_coordinates = {f"{phrase}": v for phrase, v in zip(phrases, xywh)}
    return annotated_frame, label_coordinates


def predict(model, image, caption, box_threshold, text_threshold):