一、導包、設定全局變量
import os
import sys
import time
import numpy as np
import imgaug # https://github.com/aleju/imgaug (pip3 install imgaug)
from pycocotools.coco import COCO
from pycocotools.cocoeval import COCOeval
from pycocotools import mask as maskUtils
import zipfile
import urllib.request
import shutil
#設定根目錄
ROOT_DIR = os.path.abspath(".")
#導入Mask RCNN
sys.path.append(ROOT_DIR)
from mrcnn.config import Config
from mrcnn import model as modellib, utils
#預訓練權重檔案路徑
COCO_MODEL_PATH = os.path.join(ROOT_DIR, "mask_rcnn_coco.h5")
#儲存logs和model checkpoints的路徑,可以通過指令行參數--logs設定
DEFAULT_LOGS_DIR = os.path.join(ROOT_DIR, "logs")
DEFAULT_DATASET_YEAR = "2014"
二、配置
class CocoConfig(Config):
""訓練MS COCO資料集的配置資訊。
從Config class基類繼承并重寫訓練COCO資料集相關的值。
"""
#為該配置命名
NAME = "coco"
#每個GPU一次處理多少幅圖像
IMAGES_PER_GPU = 1
#共有多少GPU參與訓練(預設是1)
# GPU_COUNT = 8
#有多少類物體需要分類 (包括背景)
NUM_CLASSES = 1 + 80 # COCO資料集有80類
三、資料集
3.1 加載coco資料集
class CocoDataset(utils.Dataset):
def load_coco(self, dataset_dir, subset, year=DEFAULT_DATASET_YEAR, class_ids=None,
class_map=None, return_coco=False, auto_download=False):
"""加載COCO資料集的一個子集.
dataset_dir:COCO資料集的根目錄.
subset:需要下載下傳的子集(train, val, minival, valminusminival)
year:下載下傳哪一年的資料集(2014, 2017)，字元串而非整型
class_ids:如果給出，則隻加載給定類别的圖像.
class_map: TODO: 還未實作. 将不同資料集的類别映射到相同的class ID.
return_coco: If True, returns the COCO object.
auto_download: 自動下載下傳并解壓MS-COCO圖像和标注檔案
"""
if auto_download is True:
self.auto_download(dataset_dir, subset, year)
coco = COCO("{}/annotations/instances_{}{}.json".format(dataset_dir, subset, year))
if subset == "minival" or subset == "valminusminival":
subset = "val"
image_dir = "{}/{}{}".format(dataset_dir, subset, year)
#加載全部類别還是某個子集?
if not class_ids:
#全部類别
class_ids = sorted(coco.getCatIds())
#全部圖像還是某個子集?
if class_ids:
image_ids = []
for id in class_ids:
image_ids.extend(list(coco.getImgIds(catIds=[id])))
#去掉重複
image_ids = list(set(image_ids))
else:
#全部圖像
image_ids = list(coco.imgs.keys())
#添加類别
for i in class_ids:
self.add_class("coco", i, coco.loadCats(i)[0]["name"])
#添加圖像
for i in image_ids:
self.add_image(
"coco", image_id=i,
path=os.path.join(image_dir, coco.imgs[i]['file_name']),
width=coco.imgs[i]["width"],
height=coco.imgs[i]["height"],
annotations=coco.loadAnns(coco.getAnnIds(
imgIds=[i], catIds=class_ids, iscrowd=None)))
if return_coco:
return coco
3.2 自動下載下傳資料集
def auto_download(self, dataDir, dataType, dataYear):
"""下載下傳COCO資料集/标注檔案.
dataDir:COCO資料集的根目錄.
dataType:下載下傳哪個子集(train, val, minival, valminusminival)
dataYear:下載下傳哪一年的資料(2014, 2017)，string類型而非整型
Note:
For 2014, "train", "val", "minival", or "valminusminival"
For 2017,"train" and "val"
"""
#設定路徑和檔案名
if dataType == "minival" or dataType == "valminusminival":
imgDir = "{}/{}{}".format(dataDir, "val", dataYear)
imgZipFile = "{}/{}{}.zip".format(dataDir, "val", dataYear)
imgURL = "http://images.cocodataset.org/zips/{}{}.zip".format("val", dataYear)
else:
imgDir = "{}/{}{}".format(dataDir, dataType, dataYear)
imgZipFile = "{}/{}{}.zip".format(dataDir, dataType, dataYear)
imgURL = "http://images.cocodataset.org/zips/{}{}.zip".format(dataType, dataYear)
# print("Image paths:"); print(imgDir); print(imgZipFile); print(imgURL)
#若檔案夾不存在，則建立
if not os.path.exists(dataDir):
os.makedirs(dataDir)
#如果圖像不存在，則下載下傳
if not os.path.exists(imgDir):
os.makedirs(imgDir)
print("Downloading images to " + imgZipFile + " ...")
with urllib.request.urlopen(imgURL) as resp, open(imgZipFile, 'wb') as out:
shutil.copyfileobj(resp, out)
print("... done downloading.")
print("Unzipping " + imgZipFile)
with zipfile.ZipFile(imgZipFile, "r") as zip_ref:
zip_ref.extractall(dataDir)
print("... done unzipping")
print("Will use images in " + imgDir)
#設定标注檔案路徑
annDir = "{}/annotations".format(dataDir)
if dataType == "minival":
annZipFile = "{}/instances_minival2014.json.zip".format(dataDir)
annFile = "{}/instances_minival2014.json".format(annDir)
annURL = "https://dl.dropboxusercontent.com/s/o43o90bna78omob/instances_minival2014.json.zip?dl=0"
unZipDir = annDir
elif dataType == "valminusminival":
annZipFile = "{}/instances_valminusminival2014.json.zip".format(dataDir)
annFile = "{}/instances_valminusminival2014.json".format(annDir)
annURL = "https://dl.dropboxusercontent.com/s/s3tw5zcg7395368/instances_valminusminival2014.json.zip?dl=0"
unZipDir = annDir
else:
annZipFile = "{}/annotations_trainval{}.zip".format(dataDir, dataYear)
annFile = "{}/instances_{}{}.json".format(annDir, dataType, dataYear)
annURL = "http://images.cocodataset.org/annotations/annotations_trainval{}.zip".format(dataYear)
unZipDir = dataDir
# print("Annotations paths:"); print(annDir); print(annFile); print(annZipFile); print(annURL)
#如果标注檔案不存在，則下載下傳
if not os.path.exists(annDir):
os.makedirs(annDir)
if not os.path.exists(annFile):
if not os.path.exists(annZipFile):
print("Downloading zipped annotations to " + annZipFile + " ...")
with urllib.request.urlopen(annURL) as resp, open(annZipFile, 'wb') as out:
shutil.copyfileobj(resp, out)
print("... done downloading.")
print("Unzipping " + annZipFile)
with zipfile.ZipFile(annZipFile, "r") as zip_ref:
zip_ref.extractall(unZipDir)
print("... done unzipping")
print("Will use annotations in " + annFile)
3.3 加載mask
def load_mask(self, image_id):
"""加載給定圖像的instance masks.
不同的資料集使用不同的方式存儲masks。本函數将不同格式的mask轉化
成同一格式的bitmap，次元是[height, width, instances].
傳回:
masks:一個bool數組，尺寸是[height, width, instance count]，
一個instance有一個mask.
class_ids:一個一維數組，元素是instance masks的class IDs。
"""
#如果不是COCO資料集的圖像, 則退化到其父類.
image_info = self.image_info[image_id]
if image_info["source"] != "coco":
return super(CocoDataset, self).load_mask(image_id)
instance_masks = []
class_ids = []
annotations = self.image_info[image_id]["annotations"]
#建立次元是[height, width, instance_count]的mask和
#每個通道的mask對應的class IDs的清單。
for annotation in annotations:
class_id = self.map_source_class_id(
"coco.{}".format(annotation['category_id']))
if class_id:
m = self.annToMask(annotation, image_info["height"],
image_info["width"])
#有些物體的面積小于一個像素，跳過這些物體，不予處理
if m.max() < 1:
continue
#是否是crowd?如果是，則賦給一個negative class ID.
if annotation['iscrowd']:
#crowds的使用negative class ID
class_id *= -1
#對于crowd masks, annToMask()傳回的mask
#有時候會小于給定的尺寸。如果是的話，将其縮放。
if m.shape[0] != image_info["height"] or m.shape[1] != image_info["width"]:
m = np.ones([image_info["height"], image_info["width"]], dtype=bool)
instance_masks.append(m)
class_ids.append(class_id)
#将instance masks打包成數組
if class_ids:
mask = np.stack(instance_masks, axis=2).astype(np.bool)
class_ids = np.array(class_ids, dtype=np.int32)
return mask, class_ids
else:
#調用super class傳回一個空的mask
return super(CocoDataset, self).load_mask(image_id)
def image_reference(self, image_id):
"""傳回圖像在COCO Website上的連結"""
info = self.image_info[image_id]
if info["source"] == "coco":
return "http://cocodataset.org/#explore?id={}".format(info["id"])
else:
super(CocoDataset, self).image_reference(image_id)
#下面的兩個函數來自pycocotools，并做了一些小的修改。
def annToRLE(self, ann, height, width):
"""
轉化多邊形的标注, 未壓縮的RLE轉成RLE.
:傳回:二值mask (numpy 2D array)
"""
segm = ann['segmentation']
if isinstance(segm, list):
# 多邊形 -- 一個物體可能由多個部分構成
# 将所有部分合并到一個mask rle code
rles = maskUtils.frPyObjects(segm, height, width)
rle = maskUtils.merge(rles)
elif isinstance(segm['counts'], list):
#未壓縮的RLE
rle = maskUtils.frPyObjects(segm, height, width)
else:
# rle
rle = ann['segmentation']
return rle
def annToMask(self, ann, height, width):
"""
轉化多邊形的标注, 未壓縮的RLE或者RLE轉成二值的mask.
:傳回: 二值的mask (numpy 2D array)
"""
rle = self.annToRLE(ann, height, width)
m = maskUtils.decode(rle)
return m
四、COCO評估
def build_coco_results(dataset, image_ids, rois, class_ids, scores, masks):
"""排列結果以符合COCO的說明，http://cocodataset.org/#format
"""
#如果沒有結果，則傳回一個空清單
if rois is None:
return []
results = []
for image_id in image_ids:
#循環擷取檢測結果
for i in range(rois.shape[0]):
class_id = class_ids[i]
score = scores[i]
bbox = np.around(rois[i], 1)
mask = masks[:, :, i]
result = {
"image_id": image_id,
"category_id": dataset.get_source_class_id(class_id, "coco"),
"bbox": [bbox[1], bbox[0], bbox[3] - bbox[1], bbox[2] - bbox[0]],
"score": score,
"segmentation": maskUtils.encode(np.asfortranarray(mask))
}
results.append(result)
return results
def evaluate_coco(model, dataset, coco, eval_type="bbox", limit=0, image_ids=None):
"""運作官方的COCO驗證
dataset:驗證資料集
eval_type: "bbox"對應bounding box，"segm"對應segmentation evaluation
limit:如果非0，則使用全部的驗證圖像
"""
#從資料集中選擇COCO圖像
image_ids = image_ids or dataset.image_ids
#限制使用多少圖像
if limit:
image_ids = image_ids[:limit]
#獲得對應的COCO image IDs.
coco_image_ids = [dataset.image_info[id]["id"] for id in image_ids]
t_prediction = 0
t_start = time.time()
results = []
for i, image_id in enumerate(image_ids):
# Load image
image = dataset.load_image(image_id)
#開始檢測
t = time.time()
r = model.detect([image], verbose=0)[0]
t_prediction += (time.time() - t)
#将結果轉化成COCO的歌聲
# 将masks轉成uint8類型，因為使用bool型COCO tools會報錯
image_results = build_coco_results(dataset, coco_image_ids[i:i + 1],
r["rois"], r["class_ids"],
r["scores"],
r["masks"].astype(np.uint8))
results.extend(image_results)
#加載結果.會用一些附加的屬性對其做一些修改.
coco_results = coco.loadRes(results)
#評估
cocoEval = COCOeval(coco, coco_results, eval_type)
cocoEval.params.imgIds = coco_image_ids
cocoEval.evaluate()
cocoEval.accumulate()
cocoEval.summarize()
print("Prediction time: {}. Average {}/image".format(
t_prediction, t_prediction / len(image_ids)))
print("Total time: ", time.time() - t_start)
五、訓練
if __name__ == '__main__':
import argparse
#解析指令行參數
parser = argparse.ArgumentParser(
description='Train Mask R-CNN on MS COCO.')
parser.add_argument("command",
metavar="<command>",
help="'train' or 'evaluate' on MS COCO")
parser.add_argument('--dataset', required=True,
metavar="/path/to/coco/",
help='Directory of the MS-COCO dataset')
parser.add_argument('--year', required=False,
default=DEFAULT_DATASET_YEAR,
metavar="<year>",
help='Year of the MS-COCO dataset (2014 or 2017) (default=2014)')
parser.add_argument('--model', 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('--limit', required=False,
default=500,
metavar="<image count>",
help='Images to use for evaluation (default=500)')
parser.add_argument('--download', required=False,
default=False,
metavar="<True|False>",
help='Automatically download and unzip MS-COCO files (default=False)',
type=bool)
args = parser.parse_args()
print("Command: ", args.command)
print("Model: ", args.model)
print("Dataset: ", args.dataset)
print("Year: ", args.year)
print("Logs: ", args.logs)
print("Auto Download: ", args.download)
#配置
if args.command == "train":
config = CocoConfig()
else:
class InferenceConfig(CocoConfig):
#因為在預測時一次隻處理一幅圖像，是以将batch size設為1
#Batch size = GPU_COUNT * IMAGES_PER_GPU
GPU_COUNT = 1
IMAGES_PER_GPU = 1
DETECTION_MIN_CONFIDENCE = 0
config = InferenceConfig()
config.display()
#建立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)
#選擇加載權重檔案
if args.model.lower() == "coco":
model_path = COCO_MODEL_PATH
elif args.model.lower() == "last":
#最近訓練的權重
model_path = model.find_last()
elif args.model.lower() == "imagenet":
#ImageNet的權重
model_path = model.get_imagenet_weights()
else:
model_path = args.model
#加載權重
print("Loading weights ", model_path)
model.load_weights(model_path, by_name=True)
#訓練或評估
if args.command == "train":
#訓練資料集. 使用訓練集合35K的驗證集，和Mask RCNN論文一緻。
dataset_train = CocoDataset()
dataset_train.load_coco(args.dataset, "train", year=args.year, auto_download=args.download)
if args.year in '2014':
dataset_train.load_coco(args.dataset, "valminusminival", year=args.year, auto_download=args.download)
dataset_train.prepare()
#驗證資料集
dataset_val = CocoDataset()
val_type = "val" if args.year in '2017' else "minival"
dataset_val.load_coco(args.dataset, val_type, year=args.year, auto_download=args.download)
dataset_val.prepare()
#圖像增強
#以50%的機率左右翻轉圖像
augmentation = imgaug.augmenters.Fliplr(0.5)
#訓練 - Stage 1
print("Training network heads")
model.train(dataset_train, dataset_val,
learning_rate=config.LEARNING_RATE,
epochs=40,
layers='heads',
augmentation=augmentation)
#訓練 - Stage 2
# Finetune 四層以後的layers
print("Fine tune Resnet stage 4 and up")
model.train(dataset_train, dataset_val,
learning_rate=config.LEARNING_RATE,
epochs=120,
layers='4+',
augmentation=augmentation)
#訓練 - Stage 3
# Fine tune所有layers
print("Fine tune all layers")
model.train(dataset_train, dataset_val,
learning_rate=config.LEARNING_RATE / 10,
epochs=160,
layers='all',
augmentation=augmentation)
elif args.command == "evaluate":
# 驗證集
dataset_val = CocoDataset()
val_type = "val" if args.year in '2017' else "minival"
coco = dataset_val.load_coco(args.dataset, val_type, year=args.year, return_coco=True, auto_download=args.download)
dataset_val.prepare()
print("Running COCO evaluation on {} images.".format(args.limit))
evaluate_coco(model, dataset_val, coco, "bbox", limit=int(args.limit))
else:
print("'{}' is not recognized. "
"Use 'train' or 'evaluate'".format(args.command))
![簡單文檔分類——樸素貝葉斯算法樸素貝葉斯算法簡單文檔分類執行個體步驟總結樸素貝葉斯分類調用(sklearn)[圖]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)