这几天在接触mask-rcnn的原理和检测数据，包括数据集的制作和标注，在一块1080ti上训练模型，到目前位置还有问题没有解决，在matterport/mask-rcnn的github仓库的issues下找解决方案，没有找到能够解决我的问题的回答。
我的问题的在balloon案例上做修改，对标记多个class的数据进行识别和实例分割，balloon是对一个class做检测，我也尝试了只对我的一个类别进行训练，是能够训练，并得到检测结果的，但是一般的应用都是需要进行多个类别的检测和识别，街上的行人，汽车，障碍物，交通灯等等都是无人驾驶的检测对象，如果我们尝试的视觉解决方案采用mask-rcnn做实验，就必要要进行多个类别的检测，所以必须探究多个class的实例分割，下面将相关的代码放上来，作为参考使用。
训练代码
#Mask R-CNN
import os
import sys
import json
import datetime
import numpy as np
import skimage.draw
from skimage import *
# Root directory of the project
ROOT_DIR = os.path.abspath("../../")
# Import Mask RCNN
sys.path.append(ROOT_DIR) # To find local version of the library
from mrcnn.config import Config
from mrcnn import model as modellib, utils
# Path to trained weights file
COCO_WEIGHTS_PATH = os.path.join(ROOT_DIR, "mask_rcnn_coco.h5")
# Directory to save logs and model checkpoints, if not provided
# through the command line argument --logs
DEFAULT_LOGS_DIR = os.path.join(ROOT_DIR, "logs")
############################################################
# Configurations
############################################################
class DefectConfig(Config):
"""Configuration for training on the toy dataset.
Derives from the base Config class and overrides some values.
"""
# Give the configuration a recognizable name
NAME = "defect"
# We use a GPU with 12GB memory, which can fit two images.
# Adjust down if you use a smaller GPU.
IMAGES_PER_GPU = 1
# Number of classes (including background)
NUM_CLASSES = 1 + 6 # Background + defect
# Number of training steps per epoch
STEPS_PER_EPOCH = 100
# Skip detections with < 90% confidence
DETECTION_MIN_CONFIDENCE = 0.9
############################################################
# Dataset
############################################################
class DefectDataset(utils.Dataset):
def load_defect(self, dataset_dir, subset):
"""Load a subset of the Balloon dataset.
dataset_dir: Root directory of the dataset.
subset: Subset to load: train or val
"""
# Add classes. We have 6 classes to add.
self.add_class("x", 1, "a")
self.add_class("x", 2, "b")
self.add_class("x", 3, "c")
self.add_class("x", 4, "d")
self.add_class("x", 5, "e")
self.add_class("x", 6, "f")
# Train or validation dataset
assert subset in ["train", "val"]
dataset_dir = os.path.join(dataset_dir, subset)
# Load annotations
# We mostly care about the x and y coordinates of each region
# Note: In VIA 2.0, regions was changed from a dict to a list.
annotations = json.load(open(os.path.join(dataset_dir, "via_region_data.json")))
annotations = list(annotations.values()) # don't need the dict keys
# The VIA tool saves images in the JSON even if they don't have any
# annotations. Skip unannotated images.
annotations = [a for a in annotations if a['regions']]
# print("ANNOTATIONS INFO:", annotations)
idlist = []
# Add images
for a in annotations:
# Get the x, y coordinaets of points of the polygons that make up
# the outline of each object instance. These are stores in the
# shape_attributes (see json format above)
# The if condition is needed to support VIA versions 1.x and 2.x.
# print(type(a['regions'])) #list
'''if type(a['regions']) is dict:
polygons = [r['shape_attributes'] for r in a['regions'].values()]
else:
polygons = [r['shape_attributes'] for r in a['regions']]'''
#print(a['regions'])
polygons = [r['shape_attributes'] for r in a['regions']]
#name = [r['region_attributes']['type'] for r in a['regions']]
#print("[NAME INFO:]", name)
#print(type(name)) #list
'''
[NAME INFO:] [{'porosity': True}, {'porosity': True}, {'surface hollow': True},
{'porosity': True}, {'lacking of sintering': True}, {'porosity': True},
{'porosity': True}, {'porosity': True}, {'surface hollow': True},
{'surface hollow': True}, {'surface hollow': True}, {'surface hollow': True},
{'surface hollow': True}, {'surface hollow': True}, {'surface hollow': True},
{'surface hollow': True}, {'surface hollow': True}]
<class 'list'>
'''
'''name_dict = {"porosity": 1,
"crack": 2,
"porosity array": 3,
"lacking of sintering": 4,
"surface hollow": 5,
"surface scratch": 6}
name_id = [name_dict[a] for a in name]'''
class_names = [r['region_attributes'] for r in a['regions']]
for i in range(len(class_names)):
if class_names[i]["type"] == "porosity":
idlist.append(1)
elif class_names[i]["type"] == "crack":
idlist.append(2)
elif class_names[i]["type"] == "porosity":
idlist.append(3)
elif class_names[i]["type"] == "lacking of sintering":
idlist.append(4)
elif class_names[i]["type"] == "surface hollow":
idlist.append(5)
elif class_names[i]["type"] == "surface scratch":
idlist.append(6)
#i = len(name)
#for i in range(len(name)):
# name_id = [name_dict[a] for a in name[i].keys()]
from skimage.io import imread
# load_mask() needs the image size to convert polygons to masks.
# Unfortunately, VIA doesn't include it in JSON, so we must read
# the image. This is only managable since the dataset is tiny.
image_path = os.path.join(dataset_dir, a['filename'])
image = skimage.io.imread(image_path)
height, width = image.shape[:2]
self.add_image(
"defect",
image_id=a['filename'], # use file name as a unique image id
path=image_path,
class_id=np.array(idlist),
width=width, height=height,
polygons=polygons)
def load_mask(self, image_id):
"""Generate instance masks for an image.
Returns:
masks: A bool array of shape [height, width, instance count] with
one mask per instance.
class_ids: a 1D array of class IDs of the instance masks.
"""
# If not a balloon dataset image, delegate to parent class.
image_info = self.image_info[image_id]
if image_info["source"] != "defect":
return super(self.__class__, self).load_mask(image_id)
name_id = image_info["class_id"]
print("[name id]", name_id)
# Convert polygons to a bitmap mask of shape
# [height, width, instance_count]
info = self.image_info[image_id]
mask = np.zeros([info["height"], info["width"], len(info["polygons"])],
dtype=np.uint8)
class_ids = np.array(name_id, dtype=np.int32)
for i, p in enumerate(info["polygons"]):
# Get indexes of pixels inside the polygon and set them to 1
rr, cc = skimage.draw.polygon(p['all_points_y'], p['all_points_x'])
mask[rr, cc, i] = 1
# Return mask, and array of class IDs of each instance. Since we have
# one class ID only, we return an array of 1s
# return mask.astype(np.bool), np.ones([mask.shape[-1]], dtype=np.int32)
return (mask.astype(np.bool), class_ids)
def image_reference(self, image_id):
"""Return the path of the image."""
info = self.image_info[image_id]
if info["source"] == "defect":
return info["path"]
else:
super(self.__class__, self).image_reference(image_id)
def train(model):
"""Train the model."""
# Training dataset.
dataset_train = DefectDataset()
dataset_train.load_defect(args.dataset, "train")
dataset_train.prepare()
# Validation dataset
dataset_val = DefectDataset()
dataset_val.load_defect(args.dataset, "val")
dataset_val.prepare()
# *** This training schedule is an example. Update to your needs ***
# Since we're using a very small dataset, and starting from
# COCO trained weights, we don't need to train too long. Also,
# no need to train all layers, just the heads should do it.
print("Training network heads")
model.train(dataset_train, dataset_val,
learning_rate=config.LEARNING_RATE,
epochs=30,
layers='heads')
def color_splash(image, mask):
"""Apply color splash effect.
image: RGB image [height, width, 3]
mask: instance segmentation mask [height, width, instance count]
Returns result image.
"""
# Make a grayscale copy of the image. The grayscale copy still
# has 3 RGB channels, though.
gray = skimage.color.gray2rgb(skimage.color.rgb2gray(image)) * 255
# Copy color pixels from the original color image where mask is set
if mask.shape[-1] > 0:
# We're treating all instances as one, so collapse the mask into one layer
mask = (np.sum(mask, -1, keepdims=True) >= 1)
splash = np.where(mask, image, gray).astype(np.uint8)
else:
splash = gray.astype(np.uint8)
return splash
def detect_and_color_splash(model, image_path=None, video_path=None):
assert image_path or video_path
# Image or video?
if image_path:
# Run model detection and generate the color splash effect
print("Running on {}".format(args.image))
# Read image
image = skimage.io.imread(args.image)
# Detect objects
r = model.detect([image], verbose=1)[0]
# Color splash
splash = color_splash(image, r['masks'])
# Save output
file_name = "splash_{:%Y%m%dT%H%M%S}.png".format(datetime.datetime.now())
skimage.io.imsave(file_name, splash)
elif video_path:
import cv2
# Video capture
vcapture = cv2.VideoCapture(video_path)
width = int(vcapture.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(vcapture.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = vcapture.get(cv2.CAP_PROP_FPS)
# Define codec and create video writer
file_name = "splash_{:%Y%m%dT%H%M%S}.avi".format(datetime.datetime.now())
vwriter = cv2.VideoWriter(file_name,
cv2.VideoWriter_fourcc(*'MJPG'),
fps, (width, height))
count = 0
success = True
while success:
print("frame: ", count)
# Read next image
success, image = vcapture.read()
if success:
# OpenCV returns images as BGR, convert to RGB
image = image[..., ::-1]
# Detect objects
r = model.detect([image], verbose=0)[0]
# Color splash
splash = color_splash(image, r['masks'])
# RGB -> BGR to save image to video
splash = splash[..., ::-1]
# Add image to video writer
vwriter.write(splash)
count += 1
vwriter.release()
print("Saved to ", file_name)
############################################################
# Training
############################################################
if __name__ == '__main__':
import argparse
# Parse command line arguments
parser = argparse.ArgumentParser(
description='Train Mask R-CNN to detect balloons.')
parser.add_argument("command",
metavar="<command>",
help="'train' or 'splash'")
parser.add_argument('--dataset', required=False,
metavar="/path/to/balloon/dataset/",
help='Directory of the Balloon dataset')
parser.add_argument('--weights', required=True,
metavar="/path/to/weights.h5",
help="Path to weights .h5 file or 'coco'")
parser.add_argument('--logs', required=False,
default=DEFAULT_LOGS_DIR,
metavar="/path/to/logs/",
help='Logs and checkpoints directory (default=logs/)')
parser.add_argument('--image', required=False,
metavar="path or URL to image",
help='Image to apply the color splash effect on')
parser.add_argument('--video', required=False,
metavar="path or URL to video",
help='Video to apply the color splash effect on')
args = parser.parse_args()
# Validate arguments
if args.command == "train":
assert args.dataset, "Argument --dataset is required for training"
elif args.command == "splash":
assert args.image or args.video,\
"Provide --image or --video to apply color splash"
print("Weights: ", args.weights)
print("Dataset: ", args.dataset)
print("Logs: ", args.logs)
# Configurations
if args.command == "train":
config = DefectConfig()
else:
class InferenceConfig(DefectConfig):
# Set batch size to 1 since we'll be running inference on
# one image at a time. Batch size = GPU_COUNT * IMAGES_PER_GPU
GPU_COUNT = 1
IMAGES_PER_GPU = 1
config = InferenceConfig()
config.display()
# Create model
if args.command == "train":
model = modellib.MaskRCNN(mode="training", config=config,
model_dir=args.logs)
else:
model = modellib.MaskRCNN(mode="inference", config=config,
model_dir=args.logs)
# Select weights file to load
if args.weights.lower() == "coco":
weights_path = COCO_WEIGHTS_PATH
# Download weights file
if not os.path.exists(weights_path):
utils.download_trained_weights(weights_path)
elif args.weights.lower() == "last":
# Find last trained weights
weights_path = model.find_last()
elif args.weights.lower() == "imagenet":
# Start from ImageNet trained weights
weights_path = model.get_imagenet_weights()
else:
weights_path = args.weights
# Load weights
print("Loading weights ", weights_path)
if args.weights.lower() == "coco":
# Exclude the last layers because they require a matching
# number of classes
model.load_weights(weights_path, by_name=True, exclude=[
"mrcnn_class_logits", "mrcnn_bbox_fc",
"mrcnn_bbox", "mrcnn_mask"])
else:
model.load_weights(weights_path, by_name=True)
# Train or evaluate
if args.command == "train":
train(model)
elif args.command == "splash":
detect_and_color_splash(model, image_path=args.image,
video_path=args.video)
else:
print("'{}' is not recognized. "
"Use 'train' or 'splash'".format(args.command))
参考方案1
def load_student(self, dataset_dir, subset):
self.add_class("student", 1, "student")
self.add_class("student", 2, "bag")
# Train or validation dataset?
assert subset in ["train", "val"]
dataset_dir = os.path.join(dataset_dir, subset)
annotations = json.load(open(os.path.join(dataset_dir, "via_region_data.json")))
annotations = list(annotations.values()) # don't need the dict keys
annotations = [a for a in annotations if a['regions']]
# Add images
for a in annotations:
polygons = [r['shape_attributes'] for r in a['regions'].values()]
objects = [s['region_attributes'] for s in a['regions'].values()]
print(objects)
num_ids=[]
for n in objects:
#print(n)
#print(type(n))
try:
if n['object_name']=='student':
num_ids.append(1)
elif n['object_name']=='bag':
num_ids.append(2)
except:
pass
#num_ids = [int(n['object_name']) for n in objects]
image_path = os.path.join(dataset_dir, a['filename'])
image = skimage.io.imread(image_path)
height, width = image.shape[:2]
self.add_image(
"student",
image_id=a['filename'], # use file name as a unique image id
path=image_path,
width=width, height=height,
polygons=polygons,
num_ids=num_ids)参考方案2
def load_multi_number(self, dataset_dir, subset):
"""Load a subset of the number dataset.
dataset_dir: Root directory of the dataset.
subset: Subset to load: train or val
"""
# Add classes
self.add_class("object", 1, "A")
self.add_class("object", 2, "B")
self.add_class("object", 3, "C")
self.add_class("object", 4, "D")
self.add_class("object", 5, "E")
self.add_class("object", 6, "F")
self.add_class("object", 7, "G")
self.add_class("object", 8, "H")
self.add_class("object", 9, "I")
self.add_class("object", 10, "J")
self.add_class("object", 11, "K")
self.add_class("object", 12, "browl")
# Train or validation dataset?
assert subset in ["train", "val"]
dataset_dir = os.path.join(dataset_dir, subset)
annotations = json.load(open(os.path.join(dataset_dir, ".../train/via_region_data.json")))
annotations = list(annotations.values()) # don't need the dict keys
# The VIA tool saves images in the JSON even if they don't have any
# annotations. Skip unannotated images.
annotations = [a for a in annotations if a['regions']]
# Add images
for a in annotations:
# Get the x, y coordinaets of points of the polygons that make up
# the outline of each object instance. There are stores in the
# shape_attributes (see json format above)
# for b in a['regions'].values():
# polygons = [{**b['shape_attributes'], **b['region_attributes']}]
# print("string=", polygons)
# for r in a['regions'].values():
# polygons = [r['shape_attributes']]
# # print("polygons=", polygons)
# multi_numbers = [r['region_attributes']]
# print("multi_numbers=", multi_numbers)
polygons = [r['shape_attributes'] for r in a['regions'].values()]
objects = [s['region_attributes'] for s in a['regions'].values()]
# print("multi_numbers=", multi_numbers)
# num_ids = [n for n in multi_numbers['number'].values()]
# for n in multi_numbers:
num_ids = [int(n['object']) for n in objects]
# print("num_ids=", num_ids)
# print("num_ids_new=", num_ids_new)
# categories = [s['region_attributes'] for s in a['regions'].values()]
# load_mask() needs the image size to convert polygons to masks.
# Unfortunately, VIA doesn't include it in JSON, so we must read
# the image. This is only managable since the dataset is tiny.
image_path = os.path.join(dataset_dir, a['filename'])
image = skimage.io.imread(image_path)
height, width = image.shape[:2]
self.add_image(
"object",
image_id=a['filename'], # use file name as a unique image id
path=image_path,
width=width, height=height,
polygons=polygons,
num_ids=num_ids)
def load_mask(self, image_id):
"""Generate instance masks for an image.
Returns:
masks: A bool array of shape [height, width, instance count] with
one mask per instance.
class_ids: a 1D array of class IDs of the instance masks.
"""
# If not a number dataset image, delegate to parent class.
info = self.image_info[image_id]
if info["source"] != "object":
return super(self.__class__, self).load_mask(image_id)
num_ids = info['num_ids']
# Convert polygons to a bitmap mask of shape
# [height, width, instance_count]
mask = np.zeros([info["height"], info["width"], len(info["polygons"])],
dtype=np.uint8)
for i, p in enumerate(info["polygons"]):
# Get indexes of pixels inside the polygon and set them to 1
rr, cc = skimage.draw.polygon(p['all_points_y'], p['all_points_x'])
mask[rr, cc, i] = 1
# print("info['num_ids']=", info['num_ids'])
# Map class names to class IDs.
num_ids = np.array(num_ids, dtype=np.int32)
return mask, num_ids
def image_reference(self, image_id):
"""Return the path of the image."""
info = self.image_info[image_id]
if info["source"] == "object":
return info["path"]
else:
super(self.__class__, self).image_reference(image_id)