Spaces:
Running
on
Zero
Running
on
Zero
File size: 4,444 Bytes
df6c67d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 |
from typing import Dict, List, Tuple, Union
import cv2
import numpy as np
from inference.core.entities.requests.inference import (
InstanceSegmentationInferenceRequest,
KeypointsDetectionInferenceRequest,
ObjectDetectionInferenceRequest,
)
from inference.core.entities.responses.inference import (
InstanceSegmentationPrediction,
Keypoint,
KeypointsPrediction,
ObjectDetectionInferenceResponse,
ObjectDetectionPrediction,
Point,
)
from inference.core.utils.image_utils import load_image_rgb, np_image_to_base64
def draw_detection_predictions(
inference_request: Union[
ObjectDetectionInferenceRequest,
InstanceSegmentationInferenceRequest,
KeypointsDetectionInferenceRequest,
],
inference_response: Union[
ObjectDetectionInferenceResponse,
InstanceSegmentationPrediction,
KeypointsPrediction,
],
colors: Dict[str, str],
) -> bytes:
image = load_image_rgb(inference_request.image)
for box in inference_response.predictions:
color = tuple(
int(colors.get(box.class_name, "#4892EA")[i : i + 2], 16) for i in (1, 3, 5)
)
image = draw_bbox(
image=image,
box=box,
color=color,
thickness=inference_request.visualization_stroke_width,
)
if hasattr(box, "points"):
image = draw_instance_segmentation_points(
image=image,
points=box.points,
color=color,
thickness=inference_request.visualization_stroke_width,
)
if hasattr(box, "keypoints"):
draw_keypoints(
image=image,
keypoints=box.keypoints,
color=color,
thickness=inference_request.visualization_stroke_width,
)
if inference_request.visualization_labels:
image = draw_labels(
image=image,
box=box,
color=color,
)
return np_image_to_base64(image=image)
def draw_bbox(
image: np.ndarray,
box: ObjectDetectionPrediction,
color: Tuple[int, ...],
thickness: int,
) -> np.ndarray:
left_top, right_bottom = bbox_to_points(box=box)
return cv2.rectangle(
image,
left_top,
right_bottom,
color=color,
thickness=thickness,
)
def draw_instance_segmentation_points(
image: np.ndarray,
points: List[Point],
color: Tuple[int, ...],
thickness: int,
) -> np.ndarray:
points_array = np.array([(int(p.x), int(p.y)) for p in points], np.int32)
if len(points) > 2:
image = cv2.polylines(
image,
[points_array],
isClosed=True,
color=color,
thickness=thickness,
)
return image
def draw_keypoints(
image: np.ndarray,
keypoints: List[Keypoint],
color: Tuple[int, ...],
thickness: int,
) -> None:
for keypoint in keypoints:
center_coordinates = (round(keypoint.x), round(keypoint.y))
image = cv2.circle(
image,
center_coordinates,
thickness,
color,
-1,
)
def draw_labels(
image: np.ndarray,
box: Union[ObjectDetectionPrediction, InstanceSegmentationPrediction],
color: Tuple[int, ...],
) -> np.ndarray:
(x1, y1), _ = bbox_to_points(box=box)
text = f"{box.class_name} {box.confidence:.2f}"
(text_width, text_height), _ = cv2.getTextSize(
text, cv2.FONT_HERSHEY_SIMPLEX, 0.5, 1
)
button_size = (text_width + 20, text_height + 20)
button_img = np.full(
(button_size[1], button_size[0], 3), color[::-1], dtype=np.uint8
)
cv2.putText(
button_img,
text,
(10, 10 + text_height),
cv2.FONT_HERSHEY_SIMPLEX,
0.5,
(255, 255, 255),
1,
)
end_x = min(x1 + button_size[0], image.shape[1])
end_y = min(y1 + button_size[1], image.shape[0])
image[y1:end_y, x1:end_x] = button_img[: end_y - y1, : end_x - x1]
return image
def bbox_to_points(
box: Union[ObjectDetectionPrediction, InstanceSegmentationPrediction],
) -> Tuple[Tuple[int, int], Tuple[int, int]]:
x1 = int(box.x - box.width / 2)
x2 = int(box.x + box.width / 2)
y1 = int(box.y - box.height / 2)
y2 = int(box.y + box.height / 2)
return (x1, y1), (x2, y2)
|