garmentiq.landmark.refine

View Source

 1import numpy as np
 2import cv2
 3import copy
 4from garmentiq.landmark.refinement import refine_landmark_with_blur
 5from garmentiq.landmark.utils import (
 6    find_instruction_landmark_index,
 7    fill_instruction_landmark_coordinate,
 8)
 9
10
11def refine(
12    class_name: str,
13    detection_np: np.array,
14    detection_conf: np.array,
15    detection_dict: dict,
16    mask: np.array,
17    window_size: int = 5,
18    ksize: tuple = (11, 11),
19    sigmaX: float = 0.0,
20):
21    """
22    Refines detected landmarks using a blurred mask and updates the detection dictionary.
23
24    This function applies Gaussian blur to the given mask, then refines landmark coordinates
25    based on their confidence scores and local intensity structure. Only landmarks with a
26    confidence score greater than 0 are refined. The refined coordinates are used to update
27    predefined landmarks in the detection dictionary.
28
29    Args:
30        class_name (str): The name of the class to access in the detection dictionary.
31        detection_np (np.array): The initial landmark predictions. Shape: (1, N, 2).
32        detection_conf (np.array): Confidence scores for each predicted landmark. Shape: (1, N, 1).
33        detection_dict (dict): Dictionary containing landmark data for each class.
34        mask (np.array): Grayscale mask image used to guide refinement.
35        window_size (int, optional): Size of the window used in the refinement algorithm. Defaults to 5.
36        ksize (tuple, optional): Kernel size for Gaussian blur. Must be odd integers. Defaults to (11, 11).
37        sigmaX (float, optional): Gaussian kernel standard deviation in the X direction. Defaults to 0.0.
38
39    Returns:
40        tuple:
41            - refined_detection_np (np.array): Array of the same shape as `detection_np` with refined coordinates.
42            - detection_dict (dict): Updated detection dictionary with refined landmark coordinates.
43    """
44    blurred_mask = cv2.GaussianBlur(mask, ksize, sigmaX)
45
46    refined_detection_np = np.zeros(detection_np.shape)
47
48    for i, coord in enumerate(detection_np[0]):
49        if detection_conf[0, i, 0] > 0:
50            refined_x, refined_y = refine_landmark_with_blur(
51                coord[0], coord[1], blurred_mask, window_size
52            )
53            refined_detection_np[0, i] = [refined_x, refined_y]
54
55    predefined_index = find_instruction_landmark_index(
56        detection_dict[class_name]["landmarks"], predefined=True
57    )
58
59    preds = refined_detection_np[:, predefined_index, :]
60
61    detection_dict_copy = copy.deepcopy(detection_dict)
62
63    detection_dict_copy[class_name]["landmarks"] = fill_instruction_landmark_coordinate(
64        instruction_landmarks=detection_dict_copy[class_name]["landmarks"],
65        index=predefined_index,
66        fill_in_value=preds,
67    )
68
69    return refined_detection_np, detection_dict_copy

def refine( class_name: str, detection_np: <built-in function array>, detection_conf: <built-in function array>, detection_dict: dict, mask: <built-in function array>, window_size: int = 5, ksize: tuple = (11, 11), sigmaX: float = 0.0): View Source

12def refine(
13    class_name: str,
14    detection_np: np.array,
15    detection_conf: np.array,
16    detection_dict: dict,
17    mask: np.array,
18    window_size: int = 5,
19    ksize: tuple = (11, 11),
20    sigmaX: float = 0.0,
21):
22    """
23    Refines detected landmarks using a blurred mask and updates the detection dictionary.
24
25    This function applies Gaussian blur to the given mask, then refines landmark coordinates
26    based on their confidence scores and local intensity structure. Only landmarks with a
27    confidence score greater than 0 are refined. The refined coordinates are used to update
28    predefined landmarks in the detection dictionary.
29
30    Args:
31        class_name (str): The name of the class to access in the detection dictionary.
32        detection_np (np.array): The initial landmark predictions. Shape: (1, N, 2).
33        detection_conf (np.array): Confidence scores for each predicted landmark. Shape: (1, N, 1).
34        detection_dict (dict): Dictionary containing landmark data for each class.
35        mask (np.array): Grayscale mask image used to guide refinement.
36        window_size (int, optional): Size of the window used in the refinement algorithm. Defaults to 5.
37        ksize (tuple, optional): Kernel size for Gaussian blur. Must be odd integers. Defaults to (11, 11).
38        sigmaX (float, optional): Gaussian kernel standard deviation in the X direction. Defaults to 0.0.
39
40    Returns:
41        tuple:
42            - refined_detection_np (np.array): Array of the same shape as `detection_np` with refined coordinates.
43            - detection_dict (dict): Updated detection dictionary with refined landmark coordinates.
44    """
45    blurred_mask = cv2.GaussianBlur(mask, ksize, sigmaX)
46
47    refined_detection_np = np.zeros(detection_np.shape)
48
49    for i, coord in enumerate(detection_np[0]):
50        if detection_conf[0, i, 0] > 0:
51            refined_x, refined_y = refine_landmark_with_blur(
52                coord[0], coord[1], blurred_mask, window_size
53            )
54            refined_detection_np[0, i] = [refined_x, refined_y]
55
56    predefined_index = find_instruction_landmark_index(
57        detection_dict[class_name]["landmarks"], predefined=True
58    )
59
60    preds = refined_detection_np[:, predefined_index, :]
61
62    detection_dict_copy = copy.deepcopy(detection_dict)
63
64    detection_dict_copy[class_name]["landmarks"] = fill_instruction_landmark_coordinate(
65        instruction_landmarks=detection_dict_copy[class_name]["landmarks"],
66        index=predefined_index,
67        fill_in_value=preds,
68    )
69
70    return refined_detection_np, detection_dict_copy

Refines detected landmarks using a blurred mask and updates the detection dictionary.

This function applies Gaussian blur to the given mask, then refines landmark coordinates based on their confidence scores and local intensity structure. Only landmarks with a confidence score greater than 0 are refined. The refined coordinates are used to update predefined landmarks in the detection dictionary.

Arguments:

class_name (str): The name of the class to access in the detection dictionary.
detection_np (np.array): The initial landmark predictions. Shape: (1, N, 2).
detection_conf (np.array): Confidence scores for each predicted landmark. Shape: (1, N, 1).
detection_dict (dict): Dictionary containing landmark data for each class.
mask (np.array): Grayscale mask image used to guide refinement.
window_size (int, optional): Size of the window used in the refinement algorithm. Defaults to 5.
ksize (tuple, optional): Kernel size for Gaussian blur. Must be odd integers. Defaults to (11, 11).
sigmaX (float, optional): Gaussian kernel standard deviation in the X direction. Defaults to 0.0.

Returns:

tuple: - refined_detection_np (np.array): Array of the same shape as detection_np with refined coordinates. - detection_dict (dict): Updated detection dictionary with refined landmark coordinates.