Spaces:
Sleeping
Sleeping
File size: 8,775 Bytes
fc895f4 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 | """
visualizer.py
-------------
Draws segmentation results (masks, bounding boxes, labels) on floor plan images.
Useful for debugging and reviewing model predictions.
Usage:
from src.segmentation.visualizer import SegmentationVisualizer
viz = SegmentationVisualizer()
annotated = viz.draw(image, result)
viz.save(annotated, "outputs/annotated.png")
viz.show(annotated)
"""
from pathlib import Path
from typing import Optional
import cv2
import numpy as np
from .predictor import SegmentationResult, DetectedElement
# ── Colour palette (BGR for OpenCV) — one per class ───────────────────────────
CLASS_COLORS = {
"OuterWall": (50, 50, 180), # dark blue
"InnerWall": (100, 100, 220), # medium blue
"Window": (255, 200, 50), # amber
"Door": (50, 200, 200), # teal
"Stairs": (160, 80, 200), # purple
"Railing": (200, 140, 80), # tan
"Kitchen": (50, 180, 80), # green
"LivingRoom": (80, 200, 255), # sky blue
"Bedroom": (200, 100, 150), # pink
"Bathroom": (100, 220, 200), # mint
"Corridor": (180, 180, 80), # olive
"Balcony": (255, 140, 50), # orange
"Garage": (140, 140, 140), # gray
}
DEFAULT_COLOR = (200, 200, 200)
MASK_ALPHA = 0.40 # Mask transparency
TEXT_SCALE = 0.45
TEXT_THICKNESS = 1
class SegmentationVisualizer:
"""
Renders segmentation predictions over floor plan images.
Args:
mask_alpha: Opacity for filled masks (0=transparent, 1=solid).
show_boxes: Draw bounding boxes.
show_labels: Draw class name + confidence labels.
show_masks: Fill detected regions with transparent colour.
"""
def __init__(
self,
mask_alpha: float = MASK_ALPHA,
show_boxes: bool = True,
show_labels: bool = True,
show_masks: bool = True,
):
self.mask_alpha = mask_alpha
self.show_boxes = show_boxes
self.show_labels = show_labels
self.show_masks = show_masks
def draw(
self,
image: np.ndarray,
result: SegmentationResult,
min_confidence: float = 0.0,
) -> np.ndarray:
"""
Draw segmentation results on the image.
Args:
image: BGR image (H, W, 3) as numpy array.
result: SegmentationResult from FloorPlanPredictor.
min_confidence: Only draw elements above this confidence.
Returns:
Annotated BGR image.
"""
if len(image.shape) == 2:
canvas = cv2.cvtColor(image, cv2.COLOR_GRAY2BGR)
else:
canvas = image.copy()
filtered = [e for e in result.elements if e.confidence >= min_confidence]
# Draw masks first (behind everything)
if self.show_masks:
canvas = self._draw_masks(canvas, filtered)
# Draw contour outlines on top of masks
canvas = self._draw_outlines(canvas, filtered)
# Draw boxes and labels on top
for elem in filtered:
color = CLASS_COLORS.get(elem.class_name, DEFAULT_COLOR)
if self.show_boxes:
self._draw_box(canvas, elem, color)
if self.show_labels:
self._draw_label(canvas, elem, color)
# Draw legend
canvas = self._draw_legend(canvas, result)
return canvas
def draw_from_path(
self, image_path: str, result: SegmentationResult
) -> np.ndarray:
"""Load image from path and draw segmentation results."""
img = cv2.imread(image_path)
if img is None:
raise IOError(f"Could not load image: {image_path}")
return self.draw(img, result)
def save(self, image: np.ndarray, output_path: str) -> None:
"""Save annotated image to disk."""
Path(output_path).parent.mkdir(parents=True, exist_ok=True)
cv2.imwrite(output_path, image)
print(f"Saved annotation: {output_path}")
def show(self, image: np.ndarray, title: str = "Segmentation Result") -> None:
"""Display image in an OpenCV window. Press any key to close."""
cv2.imshow(title, image)
cv2.waitKey(0)
cv2.destroyAllWindows()
# ── Drawing helpers ───────────────────────────────────────────────────────
def _draw_masks(
self, canvas: np.ndarray, elements: list[DetectedElement]
) -> np.ndarray:
"""Overlay semi-transparent filled masks."""
overlay = canvas.copy()
for elem in elements:
if elem.mask is None:
continue
color = CLASS_COLORS.get(elem.class_name, DEFAULT_COLOR)
overlay[elem.mask > 0] = color
return cv2.addWeighted(overlay, self.mask_alpha, canvas, 1 - self.mask_alpha, 0)
def _draw_outlines(
self, canvas: np.ndarray, elements: list[DetectedElement]
) -> np.ndarray:
"""Draw contour outlines around each detected element's mask."""
for elem in elements:
color = CLASS_COLORS.get(elem.class_name, DEFAULT_COLOR)
# Thicker outline for walls, thinner for rooms/icons
is_wall = "Wall" in elem.class_name
thickness = 3 if is_wall else 2
if elem.mask is not None:
contours, _ = cv2.findContours(
elem.mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE
)
cv2.drawContours(canvas, contours, -1, color, thickness, cv2.LINE_AA)
elif elem.polygon:
pts = np.array(elem.polygon, dtype=np.int32).reshape((-1, 1, 2))
cv2.polylines(canvas, [pts], isClosed=True, color=color,
thickness=thickness, lineType=cv2.LINE_AA)
return canvas
def _draw_box(
self, canvas: np.ndarray, elem: DetectedElement, color: tuple
) -> None:
"""Draw bounding box rectangle."""
x1, y1, x2, y2 = elem.bbox
cv2.rectangle(canvas, (x1, y1), (x2, y2), color, thickness=2)
def _draw_label(
self, canvas: np.ndarray, elem: DetectedElement, color: tuple
) -> None:
"""Draw class name + confidence label above bounding box."""
x1, y1 = elem.bbox[0], elem.bbox[1]
label = f"{elem.class_name} {elem.confidence:.0%}"
(tw, th), _ = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, TEXT_SCALE, TEXT_THICKNESS)
# Background pill
pad = 3
cv2.rectangle(
canvas,
(x1, max(0, y1 - th - 2 * pad)),
(x1 + tw + 2 * pad, y1),
color,
thickness=-1,
)
# Text (white on coloured background)
cv2.putText(
canvas,
label,
(x1 + pad, max(th, y1 - pad)),
cv2.FONT_HERSHEY_SIMPLEX,
TEXT_SCALE,
(255, 255, 255),
TEXT_THICKNESS,
cv2.LINE_AA,
)
def _draw_legend(
self, canvas: np.ndarray, result: SegmentationResult
) -> np.ndarray:
"""Draw a class legend in the bottom-left corner."""
seen_classes = sorted(set(e.class_name for e in result.elements))
if not seen_classes:
return canvas
box_size = 14
padding = 6
line_height = box_size + padding
legend_w = 160
legend_h = len(seen_classes) * line_height + padding * 2
h, w = canvas.shape[:2]
x0 = padding
y0 = h - legend_h - padding
# Semi-transparent legend background
overlay = canvas.copy()
cv2.rectangle(overlay, (x0, y0), (x0 + legend_w, h - padding), (30, 30, 30), -1)
canvas = cv2.addWeighted(overlay, 0.7, canvas, 0.3, 0)
for i, cls_name in enumerate(seen_classes):
color = CLASS_COLORS.get(cls_name, DEFAULT_COLOR)
y = y0 + padding + i * line_height
cv2.rectangle(canvas, (x0 + padding, y), (x0 + padding + box_size, y + box_size), color, -1)
cv2.putText(
canvas,
cls_name,
(x0 + padding + box_size + 6, y + box_size - 2),
cv2.FONT_HERSHEY_SIMPLEX,
0.38,
(220, 220, 220),
1,
cv2.LINE_AA,
)
return canvas
|