def unmold_detections(self, detections, mrcnn_mask, image_shape, window):
"""Reformats the detections of one image from the format of the neural
network output to a format suitable for use in the rest of the
application.
detections: [N, (y1, x1, y2, x2, class_id, score)]
mrcnn_mask: [N, height, width, num_classes]
image_shape: [height, width, depth] Original size of the image before resizing
window: [y1, x1, y2, x2] Box in the image where the real image is
excluding the padding.
Returns:
boxes: [N, (y1, x1, y2, x2)] Bounding boxes in pixels
class_ids: [N] Integer class IDs for each bounding box
scores: [N] Float probability scores of the class_id
masks: [height, width, num_instances] Instance masks
"""
# How many detections do we have?
# Detections array is padded with zeros. Find the first class_id == 0.
zero_ix = np.where(detections[:, 4] == 0)[0]
N = zero_ix[0] if zero_ix.shape[0] > 0 else detections.shape[0]
# Extract boxes, class_ids, scores, and class-specific masks
boxes = detections[:N, :4]
class_ids = detections[:N, 4].astype(np.int32)
scores = detections[:N, 5]
masks = mrcnn_mask[np.arange(N), :, :, class_ids]
# Compute scale and shift to translate coordinates to image domain.
h_scale = image_shape[0] / (window[2] - window[0])
w_scale = image_shape[1] / (window[3] - window[1])
scale = min(h_scale, w_scale)
shift = window[:2] # y, x
scales = np.array([scale, scale, scale, scale])
shifts = np.array([shift[0], shift[1], shift[0], shift[1]])
# Translate bounding boxes to image domain
boxes = np.multiply(boxes - shifts, scales).astype(np.int32)
# Filter out detections with zero area. Often only happens in early
# stages of training when the network weights are still a bit random.
exclude_ix = np.where((boxes[:, 2] - boxes[:, 0]) *
(boxes[:, 3] - boxes[:, 1]) <= 0)[0]
if exclude_ix.shape[0] > 0:
boxes = np.delete(boxes, exclude_ix, axis=0)
class_ids = np.delete(class_ids, exclude_ix, axis=0)
scores = np.delete(scores, exclude_ix, axis=0)
masks = np.delete(masks, exclude_ix, axis=0)
N = class_ids.shape[0]
# Resize masks to original image size and set boundary threshold.
full_masks = []
for i in range(N):
# Convert neural network mask to full size mask
full_mask = utils.unmold_mask(masks[i], boxes[i], image_shape)
full_masks.append(full_mask)
full_masks = np.stack(full_masks, axis=-1)\
if full_masks else np.empty((0,) + masks.shape[1:3])
return boxes, class_ids, scores, full_masks
def draw_rois(image, rois, refined_rois, mask, class_ids, class_names, limit=10):
"""
anchors: [n, (y1, x1, y2, x2)] list of anchors in image coordinates.
proposals: [n, 4] the same anchors but refined to fit objects better.
"""
masked_image = image.copy()
# Pick random anchors in case there are too many.
ids = np.arange(rois.shape[0], dtype=np.int32)
ids = np.random.choice(
ids, limit, replace=False) if ids.shape[0] > limit else ids
fig, ax = plt.subplots(1, figsize=(12, 12))
if rois.shape[0] > limit:
plt.title("Showing {} random ROIs out of {}".format(
len(ids), rois.shape[0]))
else:
plt.title("{} ROIs".format(len(ids)))
# Show area outside image boundaries.
ax.set_ylim(image.shape[0] + 20, -20)
ax.set_xlim(-50, image.shape[1] + 20)
ax.axis('off')
for i, id in enumerate(ids):
color = np.random.rand(3)
class_id = class_ids[id]
# ROI
y1, x1, y2, x2 = rois[id]
p = patches.Rectangle((x1, y1), x2 - x1, y2 - y1, linewidth=2,
edgecolor=color if class_id else "gray",
facecolor='none', linestyle="dashed")
ax.add_patch(p)
# Refined ROI
if class_id:
ry1, rx1, ry2, rx2 = refined_rois[id]
p = patches.Rectangle((rx1, ry1), rx2 - rx1, ry2 - ry1, linewidth=2,
edgecolor=color, facecolor='none')
ax.add_patch(p)
# Connect the top-left corners of the anchor and proposal for easy
# visualization
ax.add_line(lines.Line2D([x1, rx1], [y1, ry1], color=color))
# Label
label = class_names[class_id]
ax.text(rx1, ry1 + 8, "{}".format(label),
color='w', size=11, backgroundcolor="none")
# Mask
m = utils.unmold_mask(mask[id], rois[id]
[:4].astype(np.int32), image.shape)
masked_image = apply_mask(masked_image, m, color)
ax.imshow(masked_image)
# Print stats
print("Positive ROIs: ", class_ids[class_ids > 0].shape[0])
print("Negative ROIs: ", class_ids[class_ids == 0].shape[0])
print("Positive Ratio: {:.2f}".format(
class_ids[class_ids > 0].shape[0] / class_ids.shape[0]))
def draw_rois(image, rois, refined_rois, mask, class_ids, class_names, limit=10):
"""
anchors: [n, (y1, x1, y2, x2)] list of anchors in image coordinates.
proposals: [n, 4] the same anchors but refined to fit objects better.
"""
masked_image = image.copy()
# Pick random anchors in case there are too many.
ids = np.arange(rois.shape[0], dtype=np.int32)
ids = np.random.choice(
ids, limit, replace=False) if ids.shape[0] > limit else ids
fig, ax = plt.subplots(1, figsize=(12, 12))
if rois.shape[0] > limit:
plt.title("Showing {} random ROIs out of {}".format(
len(ids), rois.shape[0]))
else:
plt.title("{} ROIs".format(len(ids)))
# Show area outside image boundaries.
ax.set_ylim(image.shape[0] + 20, -20)
ax.set_xlim(-50, image.shape[1] + 20)
ax.axis('off')
for i, id in enumerate(ids):
color = np.random.rand(3)
class_id = class_ids[id]
# ROI
y1, x1, y2, x2 = rois[id]
p = patches.Rectangle((x1, y1), x2 - x1, y2 - y1, linewidth=2,
edgecolor=color if class_id else "gray",
facecolor='none', linestyle="dashed")
ax.add_patch(p)
# Refined ROI
if class_id:
ry1, rx1, ry2, rx2 = refined_rois[id]
p = patches.Rectangle((rx1, ry1), rx2 - rx1, ry2 - ry1, linewidth=2,
edgecolor=color, facecolor='none')
ax.add_patch(p)
# Connect the top-left corners of the anchor and proposal for easy visualization
ax.add_line(lines.Line2D([x1, rx1], [y1, ry1], color=color))
# Label
label = class_names[class_id]
ax.text(rx1, ry1 + 8, "{}".format(label),
color='w', size=11, backgroundcolor="none")
# Mask
m = utils.unmold_mask(mask[id], rois[id]
[:4].astype(np.int32), image.shape)
masked_image = apply_mask(masked_image, m, color)
ax.imshow(masked_image)
# Print stats
print("Positive ROIs: ", class_ids[class_ids > 0].shape[0])
print("Negative ROIs: ", class_ids[class_ids == 0].shape[0])
print("Positive Ratio: {:.2f}".format(
class_ids[class_ids > 0].shape[0] / class_ids.shape[0]))
def draw_rois(image, rois, refined_rois, mask, class_ids, class_names, limit=10):
masked_image = image.copy()
# Выбрать случайные рамки на случай, если их слишком много.
ids = np.arange(rois.shape[0], dtype=np.int32)
ids = np.random.choice(
ids, limit, replace=False) if ids.shape[0] > limit else ids
fig, ax = plt.subplots(1, figsize=(12, 12))
if rois.shape[0] > limit:
plt.title("Showing {} random ROIs out of {}".format(
len(ids), rois.shape[0]))
else:
plt.title("{} ROIs".format(len(ids)))
# Показать область за пределами изображения.
ax.set_ylim(image.shape[0] + 20, -20)
ax.set_xlim(-50, image.shape[1] + 20)
ax.axis('off')
for i, id in enumerate(ids):
color = np.random.rand(3)
class_id = class_ids[id]
# ROI
y1, x1, y2, x2 = rois[id]
p = patches.Rectangle((x1, y1), x2 - x1, y2 - y1, linewidth=2,
edgecolor=color if class_id else "gray",
facecolor='none', linestyle="dashed")
ax.add_patch(p)
# Уточненный ROI
if class_id:
ry1, rx1, ry2, rx2 = refined_rois[id]
p = patches.Rectangle((rx1, ry1), rx2 - rx1, ry2 - ry1, linewidth=2, edgecolor=color, facecolor='none')
ax.add_patch(p)
# Соединить верхние левые углы рамки и предсказанного
# для легкой визуализации
ax.add_line(lines.Line2D([x1, rx1], [y1, ry1], color=color))
# Метки
label = class_names[class_id]
ax.text(rx1, ry1 + 8, "{}".format(label), color='w', size=11, backgroundcolor="none")
# Маски
m = utils.unmold_mask(mask[id], rois[id][:4].astype(np.int32), image.shape)
masked_image = apply_mask(masked_image, m, color)
ax.imshow(masked_image)
# Печать статистики
print("Positive ROIs: ", class_ids[class_ids > 0].shape[0])
print("Negative ROIs: ", class_ids[class_ids == 0].shape[0])
print("Positive Ratio: {:.2f}".format(
class_ids[class_ids > 0].shape[0] / class_ids.shape[0]))
# Get detection class IDs. Trim zero padding.
det_class_ids = mrcnn['detections'][0, :, 4].astype(np.int32)
det_count = np.where(det_class_ids == 0)[0][0]
det_class_ids = det_class_ids[:det_count]
print("{} detections: {}".format(det_count,
np.array(dataset.class_names)[det_class_ids]))
# In[29]:
# Masks
det_boxes = mrcnn["detections"][0, :, :4].astype(np.int32)
det_mask_specific = np.array(
[mrcnn["masks"][0, i, :, :, c] for i, c in enumerate(det_class_ids)])
det_masks = np.array([
utils.unmold_mask(m, det_boxes[i], image.shape)
for i, m in enumerate(det_mask_specific)
])
log("det_mask_specific", det_mask_specific)
log("det_masks", det_masks)
# In[30]:
display_images(det_mask_specific[:4] * 255, cmap="Blues", interpolation="none")
# In[31]:
display_images(det_masks[:4] * 255, cmap="Blues", interpolation="none")
# ## Visualize Activations
#
def unmold_detections(detections, mrcnn_mask, original_image_shape,
image_shape, window):
"""Reformats the detections of one image from the format of the neural
network output to a format suitable for use in the rest of the
application.
detections: [N, (y1, x1, y2, x2, class_id, score)] in normalized coordinates
mrcnn_mask: [N, height, width, num_classes]
original_image_shape: [H, W, C] Original image shape before resizing
image_shape: [H, W, C] Shape of the image after resizing and padding
window: [y1, x1, y2, x2] Pixel coordinates of box in the image where the real
image is excluding the padding.
Returns:
boxes: [N, (y1, x1, y2, x2)] Bounding boxes in pixels
class_ids: [N] Integer class IDs for each bounding box
scores: [N] Float probability scores of the class_id
masks: [height, width, num_instances] Instance masks
"""
# How many detections do we have?
# Detections array is padded with zeros. Find the first class_id == 0.
zero_ix = np.where(detections[:, 4] == 0)[0]
N = zero_ix[0] if zero_ix.shape[0] > 0 else detections.shape[0]
# Extract boxes, class_ids, scores, and class-specific masks
boxes = detections[:N, :4]
class_ids = detections[:N, 4].astype(np.int32)
scores = detections[:N, 5]
masks = mrcnn_mask[np.arange(N), :, :, class_ids]
# Translate normalized coordinates in the resized image to pixel
# coordinates in the original image before resizing
window = utils.norm_boxes(window, image_shape[:2])
wy1, wx1, wy2, wx2 = window
shift = np.array([wy1, wx1, wy1, wx1])
wh = wy2 - wy1 # window height
ww = wx2 - wx1 # window width
scale = np.array([wh, ww, wh, ww])
# Convert boxes to normalized coordinates on the window
boxes = np.divide(boxes - shift, scale)
# Convert boxes to pixel coordinates on the original image
boxes = utils.denorm_boxes(boxes, original_image_shape[:2])
# Filter out detections with zero area. Happens in early training when
# network weights are still random
exclude_ix = np.where((boxes[:, 2] - boxes[:, 0]) *
(boxes[:, 3] - boxes[:, 1]) <= 0)[0]
if exclude_ix.shape[0] > 0:
boxes = np.delete(boxes, exclude_ix, axis=0)
class_ids = np.delete(class_ids, exclude_ix, axis=0)
scores = np.delete(scores, exclude_ix, axis=0)
masks = np.delete(masks, exclude_ix, axis=0)
N = class_ids.shape[0]
# Resize masks to original image size and set boundary threshold.
full_masks = []
for i in range(N):
# Convert neural network mask to full size mask
full_mask = utils.unmold_mask(masks[i], boxes[i], original_image_shape)
full_masks.append(full_mask)
full_masks = np.stack(full_masks, axis=-1)\
if full_masks else np.empty(masks.shape[1:3] + (0,))
return boxes, class_ids, scores, full_masks
det_class_ids = mrcnn['detections'][0, :, 4].astype(np.int32)
det_count = np.where(det_class_ids == 0)[0][0]
det_class_ids = det_class_ids[:det_count]
print("{} detections: {}".format(
det_count, np.array(dataset.class_names)[det_class_ids]))
######################################
# Masks
det_boxes = mrcnn["detections"][0, :, :4].astype(np.int32)
det_mask_specific = np.array([mrcnn["masks"][0, i, :, :, c]
for i, c in enumerate(det_class_ids)])
det_masks = np.array([utils.unmold_mask(m, det_boxes[i], image.shape)
for i, m in enumerate(det_mask_specific)])
log("det_mask_specific", det_mask_specific)
log("det_masks", det_masks)
########################################
display_images(det_mask_specific[:4] * 255, cmap="Blues", interpolation="none")
######################################
("masks", model.keras_model.get_layer("mrcnn_mask").output),
])
# Get detection class IDs. Trim zero padding.
det_class_ids = mrcnn['detections'][0, :, 4].astype(np.int32)
det_count = np.where(det_class_ids == 0)[0][0]
det_class_ids = det_class_ids[:det_count]
print("{} detections: {}".format(
det_count, np.array(class_names)[det_class_ids]))
# Masks
det_boxes = mrcnn["detections"][0, :, :4].astype(np.int32)
det_mask_specific = np.array([mrcnn["masks"][0, i, :, :, c]
for i, c in enumerate(det_class_ids)])
det_masks = np.array([utils.unmold_mask(m, det_boxes[i], image.shape)
for i, m in enumerate(det_mask_specific)])
log("det_mask_specific", det_mask_specific)
log("det_masks", det_masks)
display_images(det_mask_specific[:4] * 255, cmap="Blues", interpolation="none")
display_images(det_masks[:4] * 255, cmap="Blues", interpolation="none")
# Get activations of a few sample layers
activations = model.run_graph([image], [
("input_image", model.keras_model.get_layer("input_image").output),
("res4w_out", model.keras_model.get_layer("res4w_out").output), # for resnet100
("rpn_bbox", model.keras_model.get_layer("rpn_bbox").output),
("roi", model.keras_model.get_layer("ROI").output),
])
def get_predicted_masks(dataset, image_id, model_pred_data_path):
# ********** Model Predictions
data_path = model_pred_data_path + image_id
pred = load_data_single(data_path)
ranks = np.squeeze(pred, axis=0)
og_image = dataset.load_image(image_id)
original_shape = og_image.shape
# Pre-processed Object Mask
pre_proc_data = dataset.load_obj_pre_proc_data(image_id)
obj_masks = pre_proc_data["obj_masks"]
obj_masks = np.squeeze(obj_masks, axis=0)
# Get Original bounding box for resize
object_roi_masks = dataset.load_object_roi_masks(image_id)
obj_roi = utils.extract_bboxes(object_roi_masks)
# ********** Process Predicted Masks
# First get saliency values
masks = np.array_split(obj_masks, 2, axis=-1)
masks = masks[1]
masks = np.squeeze(masks, axis=-1)
# Remove "padded" data
masks = masks[:obj_roi.shape[0]]
ranks = ranks[:obj_roi.shape[0]]
# Resize the predicted masks to original size and location based on their available bounding boxes
# Resize masks to original image size and set boundary threshold.
N = obj_roi.shape[0]
full_masks = []
for j in range(N):
# Convert neural network mask to full size mask
full_mask = utils.unmold_mask(masks[j], obj_roi[j], original_shape)
full_masks.append(full_mask)
full_masks = np.stack(full_masks, axis=-1) \
if full_masks else np.empty(original_shape[:2] + (0,))
# ********** Process Predicted Ranks
sorted_ranks, supposed_ranks = get_ranks_infer(ranks)
num_positive_pred = sum(r > 0 for r in supposed_ranks)
num_pred_ranks = len(sorted_ranks)
sal_seg_maps_list = []
for j in range(num_pred_ranks):
p_r = sorted_ranks[j][0]
idx = sorted_ranks[j][1]
# Break once we reach BG objects
if not p_r > 0:
break
sal_map = full_masks[:, :, idx]
sal_seg_maps_list.append(sal_map)
predicted_masks = None
has_prediction = False
if num_positive_pred > 0:
predicted_masks = np.stack(sal_seg_maps_list)
has_prediction = True
return predicted_masks, has_prediction
def display_instances(image,
boxes,
masks,
class_ids,
class_names,
scores=None,
title="",
figsize=(16, 16),
ax=None):
"""
boxes: [num_instance, (y1, x1, y2, x2, class_id)] in image coordinates.
masks: [num_instances, height, width]
class_ids: [num_instances]
class_names: list of class names of the dataset
scores: (optional) confidence scores for each box
figsize: (optional) the size of the image.
"""
# Number of instances
N = boxes.shape[0]
if not N:
print("\n*** No instances to display *** \n")
else:
assert boxes.shape[0] == masks.shape[0] == class_ids.shape[0]
if not ax:
_, ax = plt.subplots(1, figsize=figsize)
# Generate random colors
colors = random_colors(N)
# Show area outside image boundaries.
height, width = image.shape[:2]
ax.set_ylim(height + 10, -10)
ax.set_xlim(-10, width + 10)
ax.axis('off')
ax.set_title(title)
masked_image = image.astype(np.uint32).copy()
for i in range(N):
color = colors[i]
# Bounding box
if not np.any(boxes[i]):
# Skip this instance. Has no bbox. Likely lost in image cropping.
continue
y1, x1, y2, x2 = boxes[i]
p = patches.Rectangle((x1, y1),
x2 - x1,
y2 - y1,
linewidth=2,
alpha=0.7,
linestyle="dashed",
edgecolor=color,
facecolor='none')
ax.add_patch(p)
# Label
class_id = class_ids[i]
score = scores[i] if scores is not None else None
label = class_names[class_id]
x = random.randint(x1, (x1 + x2) // 2)
caption = "{} {:.3f}".format(label, score) if score else label
ax.text(x1,
y1 + 8,
caption,
color='w',
size=11,
backgroundcolor="none")
# Mask
# Resize masks to original image size and set boundary threshold.
mask = utils.unmold_mask(masks[i], boxes[i], image.shape)
masked_image = apply_mask(masked_image, mask, color)
# Mask Polygon
# Pad to ensure proper polygons for masks that touch image edges.
padded_mask = np.zeros((mask.shape[0] + 2, mask.shape[1] + 2),
dtype=np.uint8)
padded_mask[1:-1, 1:-1] = mask
contours = find_contours(padded_mask, 0.5)
for verts in contours:
# Subtract the padding and flip (y, x) to (x, y)
verts = np.fliplr(verts) - 1
p = Polygon(verts, facecolor="none", edgecolor=color)
ax.add_patch(p)
ax.imshow(masked_image.astype(np.uint8))
plt.show()