Keras版Faster-RCNN代码学习（IOU，RPN）1
Keras版Faster-RCNN代码学习（Batch Normalization）2
Keras版Faster-RCNN代码学习（loss，xml解析）3
Keras版Faster-RCNN代码学习（roipooling resnet/vgg）4
Keras版Faster-RCNN代码学习（measure_map，train/test）5
mAP
mAP是目标算法中衡量算法的精确度的指标，涉及两个概念：查准率Precision、查全率Recall。对于object detection任务，每一个object都可以计算出其Precision和Recall，多次计算/试验，每个类都 可以得到一条P-R曲线，曲线下的面积就是AP的值，这个mean的意思是对每个类的AP再求平均，得到的就是mAP的值，mAP的大小一定在[0,1]区间。
AP:Precision对Recall积分，可通过改变正负样本阈值求得矩形面积，进而求积分得到，也可以通过sklearn.metrics.average_precision_score函数直接得到。
mAP:每类AP求均值
measure_py
传入预测值和真实值和resize比例，得到可以传入sklearn.metrics.average_precision_score函数的值，即：真实值和预测概率
def get_map(pred, gt, f):
T = {}
P = {}
fx, fy = f
for bbox in gt:
bbox['bbox_matched'] = False
pred_probs = np.array([s['prob'] for s in pred])
box_idx_sorted_by_prob = np.argsort(pred_probs)[::-1]
for box_idx in box_idx_sorted_by_prob:
pred_box = pred[box_idx]
pred_class = pred_box['class']
pred_x1 = pred_box['x1']
pred_x2 = pred_box['x2']
pred_y1 = pred_box['y1']
pred_y2 = pred_box['y2']
pred_prob = pred_box['prob']
if pred_class not in P:
P[pred_class] = []
T[pred_class] = []
P[pred_class].append(pred_prob)
found_match = False
for gt_box in gt:
gt_class = gt_box['class']
gt_x1 = gt_box['x1']/fx
gt_x2 = gt_box['x2']/fx
gt_y1 = gt_box['y1']/fy
gt_y2 = gt_box['y2']/fy
gt_seen = gt_box['bbox_matched']
if gt_class != pred_class:
continue
if gt_seen:
continue
iou = data_generators.iou((pred_x1, pred_y1, pred_x2, pred_y2), (gt_x1, gt_y1, gt_x2, gt_y2))
if iou >= 0.5:
found_match = True
gt_box['bbox_matched'] = True
break
else:
continue
T[pred_class].append(int(found_match))
#漏检的计入
for gt_box in gt:
if not gt_box['bbox_matched'] and not gt_box['difficult']:
if gt_box['class'] not in P:
P[gt_box['class']] = []
T[gt_box['class']] = []
T[gt_box['class']].append(1)
P[gt_box['class']].append(0)
#import pdb
#pdb.set_trace()
return T, Presize比例计算
def format_img(img, C):
img_min_side = float(C.im_size)
(height,width,_) = img.shape
if width <= height:
f = img_min_side/width
new_height = int(f * height)
new_width = int(img_min_side)
else:
f = img_min_side/height
new_width = int(f * width)
new_height = int(img_min_side)
fx = width/float(new_width)
fy = height/float(new_height)
img = cv2.resize(img, (new_width, new_height), interpolation=cv2.INTER_CUBIC)
img = img[:, :, (2, 1, 0)]
img = img.astype(np.float32)
img[:, :, 0] -= C.img_channel_mean[0]
img[:, :, 1] -= C.img_channel_mean[1]
img[:, :, 2] -= C.img_channel_mean[2]
img /= C.img_scaling_factor
img = np.transpose(img, (2, 0, 1))
img = np.expand_dims(img, axis=0)
return img, fx, fy然后搭建网络，载入权重，传递参数，计算mAP～～
train_frcnn
from __future__ import division
import random
import pprint
import sys
import time
import numpy as np
from optparse import OptionParser
import pickle
from keras import backend as K
from keras.optimizers import Adam, SGD, RMSprop
from keras.layers import Input
from keras.models import Model
from keras_frcnn import config, data_generators
from keras_frcnn import losses as losses
import keras_frcnn.roi_helpers as roi_helpers
from keras.utils import generic_utils
sys.setrecursionlimit(40000)
parser = OptionParser()
parser.add_option("-p", "--path", dest="train_path", help="Path to training data.")
parser.add_option("-o", "--parser", dest="parser", help="Parser to use. One of simple or pascal_voc",
default="pascal_voc")
parser.add_option("-n", "--num_rois", type="int", dest="num_rois", help="Number of RoIs to process at once.", default=32)
parser.add_option("--network", dest="network", help="Base network to use. Supports vgg or resnet50.", default='resnet50')
parser.add_option("--hf", dest="horizontal_flips", help="Augment with horizontal flips in training. (Default=false).", action="store_true", default=False)
parser.add_option("--vf", dest="vertical_flips", help="Augment with vertical flips in training. (Default=false).", action="store_true", default=False)
parser.add_option("--rot", "--rot_90", dest="rot_90", help="Augment with 90 degree rotations in training. (Default=false).",
action="store_true", default=False)
parser.add_option("--num_epochs", type="int", dest="num_epochs", help="Number of epochs.", default=2000)
parser.add_option("--config_filename", dest="config_filename", help=
"Location to store all the metadata related to the training (to be used when testing).",
default="config.pickle")
parser.add_option("--output_weight_path", dest="output_weight_path", help="Output path for weights.", default='./model_frcnn.hdf5')
parser.add_option("--input_weight_path", dest="input_weight_path", help="Input path for weights. If not specified, will try to load default weights provided by keras.")
(options, args) = parser.parse_args()
if not options.train_path: # if filename is not given
parser.error('Error: path to training data must be specified. Pass --path to command line')
if options.parser == 'pascal_voc':
from keras_frcnn.pascal_voc_parser import get_data
elif options.parser == 'simple':
from keras_frcnn.simple_parser import get_data
else:
raise ValueError("Command line option parser must be one of 'pascal_voc' or 'simple'")
# pass the settings from the command line, and persist them in the config object
C = config.Config()
C.use_horizontal_flips = bool(options.horizontal_flips)
C.use_vertical_flips = bool(options.vertical_flips)
C.rot_90 = bool(options.rot_90)
C.model_path = options.output_weight_path
C.num_rois = int(options.num_rois)
if options.network == 'vgg':
C.network = 'vgg'
from keras_frcnn import vgg as nn
elif options.network == 'resnet50':
from keras_frcnn import resnet as nn
C.network = 'resnet50'
else:
print('Not a valid model')
raise ValueError
# check if weight path was passed via command line
if options.input_weight_path:
C.base_net_weights = options.input_weight_path
else:
# set the path to weights based on backend and model
C.base_net_weights = nn.get_weight_path()
all_imgs, classes_count, class_mapping = get_data(options.train_path)
if 'bg' not in classes_count:
classes_count['bg'] = 0
class_mapping['bg'] = len(class_mapping)
C.class_mapping = class_mapping
inv_map = {v: k for k, v in class_mapping.items()}
print('Training images per class:')
pprint.pprint(classes_count)
print('Num classes (including bg) = {}'.format(len(classes_count)))
config_output_filename = options.config_filename
with open(config_output_filename, 'wb') as config_f:
pickle.dump(C,config_f)
print('Config has been written to {}, and can be loaded when testing to ensure correct results'.format(config_output_filename))
random.shuffle(all_imgs)
num_imgs = len(all_imgs)
train_imgs = [s for s in all_imgs if s['imageset'] == 'trainval']
val_imgs = [s for s in all_imgs if s['imageset'] == 'test']
print('Num train samples {}'.format(len(train_imgs)))
print('Num val samples {}'.format(len(val_imgs)))
data_gen_train = data_generators.get_anchor_gt(train_imgs, classes_count, C, nn.get_img_output_length, K.image_dim_ordering(), mode='train')
data_gen_val = data_generators.get_anchor_gt(val_imgs, classes_count, C, nn.get_img_output_length,K.image_dim_ordering(), mode='val')
if K.image_dim_ordering() == 'th':
input_shape_img = (3, None, None)
else:
input_shape_img = (None, None, 3)
img_input = Input(shape=input_shape_img)
roi_input = Input(shape=(None, 4))
# define the base network (resnet here, can be VGG, Inception, etc)
shared_layers = nn.nn_base(img_input, trainable=True)
# define the RPN, built on the base layers
num_anchors = len(C.anchor_box_scales) * len(C.anchor_box_ratios)
rpn = nn.rpn(shared_layers, num_anchors)
classifier = nn.classifier(shared_layers, roi_input, C.num_rois, nb_classes=len(classes_count), trainable=True)
model_rpn = Model(img_input, rpn[:2])
model_classifier = Model([img_input, roi_input], classifier)
# this is a model that holds both the RPN and the classifier, used to load/save weights for the models
model_all = Model([img_input, roi_input], rpn[:2] + classifier)
try:
print('loading weights from {}'.format(C.base_net_weights))
model_rpn.load_weights(C.base_net_weights, by_name=True)
model_classifier.load_weights(C.base_net_weights, by_name=True)
except:
print('Could not load pretrained model weights. Weights can be found in the keras application folder \
https://github.com/fchollet/keras/tree/master/keras/applications')
optimizer = Adam(lr=1e-5)
optimizer_classifier = Adam(lr=1e-5)
model_rpn.compile(optimizer=optimizer, loss=[losses.rpn_loss_cls(num_anchors), losses.rpn_loss_regr(num_anchors)])
model_classifier.compile(optimizer=optimizer_classifier, loss=[losses.class_loss_cls, losses.class_loss_regr(len(classes_count)-1)], metrics={'dense_class_{}'.format(len(classes_count)): 'accuracy'})
model_all.compile(optimizer='sgd', loss='mae')
epoch_length = 1000
num_epochs = int(options.num_epochs)
iter_num = 0
losses = np.zeros((epoch_length, 5))
rpn_accuracy_rpn_monitor = []
rpn_accuracy_for_epoch = []
start_time = time.time()
best_loss = np.Inf
class_mapping_inv = {v: k for k, v in class_mapping.items()}
print('Starting training')
vis = True
for epoch_num in range(num_epochs):
progbar = generic_utils.Progbar(epoch_length)
print('Epoch {}/{}'.format(epoch_num + 1, num_epochs))
while True:
try:
if len(rpn_accuracy_rpn_monitor) == epoch_length and C.verbose:
mean_overlapping_bboxes = float(sum(rpn_accuracy_rpn_monitor))/len(rpn_accuracy_rpn_monitor)
rpn_accuracy_rpn_monitor = []
print('Average number of overlapping bounding boxes from RPN = {} for {} previous iterations'.format(mean_overlapping_bboxes, epoch_length))
if mean_overlapping_bboxes == 0:
print('RPN is not producing bounding boxes that overlap the ground truth boxes. Check RPN settings or keep training.')
X, Y, img_data = next(data_gen_train)
loss_rpn = model_rpn.train_on_batch(X, Y)
P_rpn = model_rpn.predict_on_batch(X)
R = roi_helpers.rpn_to_roi(P_rpn[0], P_rpn[1], C, K.image_dim_ordering(), use_regr=True, overlap_thresh=0.7, max_boxes=300)
# note: calc_iou converts from (x1,y1,x2,y2) to (x,y,w,h) format
X2, Y1, Y2, IouS = roi_helpers.calc_iou(R, img_data, C, class_mapping)
if X2 is None:
rpn_accuracy_rpn_monitor.append(0)
rpn_accuracy_for_epoch.append(0)
continue
neg_samples = np.where(Y1[0, :, -1] == 1)
pos_samples = np.where(Y1[0, :, -1] == 0)
if len(neg_samples) > 0:
neg_samples = neg_samples[0]
else:
neg_samples = []
if len(pos_samples) > 0:
pos_samples = pos_samples[0]
else:
pos_samples = []
rpn_accuracy_rpn_monitor.append(len(pos_samples))
rpn_accuracy_for_epoch.append((len(pos_samples)))
if C.num_rois > 1:
if len(pos_samples) < C.num_rois//2:
selected_pos_samples = pos_samples.tolist()
else:
selected_pos_samples = np.random.choice(pos_samples, C.num_rois//2, replace=False).tolist()
try:
selected_neg_samples = np.random.choice(neg_samples, C.num_rois - len(selected_pos_samples), replace=False).tolist()
except:
selected_neg_samples = np.random.choice(neg_samples, C.num_rois - len(selected_pos_samples), replace=True).tolist()
sel_samples = selected_pos_samples + selected_neg_samples
else:
# in the extreme case where num_rois = 1, we pick a random pos or neg sample
selected_pos_samples = pos_samples.tolist()
selected_neg_samples = neg_samples.tolist()
if np.random.randint(0, 2):
sel_samples = random.choice(neg_samples)
else:
sel_samples = random.choice(pos_samples)
loss_class = model_classifier.train_on_batch([X, X2[:, sel_samples, :]], [Y1[:, sel_samples, :], Y2[:, sel_samples, :]])
losses[iter_num, 0] = loss_rpn[1]
losses[iter_num, 1] = loss_rpn[2]
losses[iter_num, 2] = loss_class[1]
losses[iter_num, 3] = loss_class[2]
losses[iter_num, 4] = loss_class[3]
iter_num += 1
progbar.update(iter_num, [('rpn_cls', np.mean(losses[:iter_num, 0])), ('rpn_regr', np.mean(losses[:iter_num, 1])),
('detector_cls', np.mean(losses[:iter_num, 2])), ('detector_regr', np.mean(losses[:iter_num, 3]))])
if iter_num == epoch_length:
loss_rpn_cls = np.mean(losses[:, 0])
loss_rpn_regr = np.mean(losses[:, 1])
loss_class_cls = np.mean(losses[:, 2])
loss_class_regr = np.mean(losses[:, 3])
class_acc = np.mean(losses[:, 4])
mean_overlapping_bboxes = float(sum(rpn_accuracy_for_epoch)) / len(rpn_accuracy_for_epoch)
rpn_accuracy_for_epoch = []
if C.verbose:
print('Mean number of bounding boxes from RPN overlapping ground truth boxes: {}'.format(mean_overlapping_bboxes))
print('Classifier accuracy for bounding boxes from RPN: {}'.format(class_acc))
print('Loss RPN classifier: {}'.format(loss_rpn_cls))
print('Loss RPN regression: {}'.format(loss_rpn_regr))
print('Loss Detector classifier: {}'.format(loss_class_cls))
print('Loss Detector regression: {}'.format(loss_class_regr))
print('Elapsed time: {}'.format(time.time() - start_time))
curr_loss = loss_rpn_cls + loss_rpn_regr + loss_class_cls + loss_class_regr
iter_num = 0
start_time = time.time()
if curr_loss < best_loss:
if C.verbose:
print('Total loss decreased from {} to {}, saving weights'.format(best_loss,curr_loss))
best_loss = curr_loss
model_all.save_weights(C.model_path)
break
except Exception as e:
print('Exception: {}'.format(e))
continue
print('Training complete, exiting.')
test_frcnn
from __future__ import division
import os
import cv2
import numpy as np
import sys
import pickle
from optparse import OptionParser
import time
from keras_frcnn import config
from keras import backend as K
from keras.layers import Input
from keras.models import Model
from keras_frcnn import roi_helpers
sys.setrecursionlimit(40000)
parser = OptionParser()
parser.add_option("-p", "--path", dest="test_path", help="Path to test data.")
parser.add_option("-n", "--num_rois", type="int", dest="num_rois",
help="Number of ROIs per iteration. Higher means more memory use.", default=32)
parser.add_option("--config_filename", dest="config_filename", help=
"Location to read the metadata related to the training (generated when training).",
default="config.pickle")
parser.add_option("--network", dest="network", help="Base network to use. Supports vgg or resnet50.", default='resnet50')
(options, args) = parser.parse_args()
if not options.test_path: # if filename is not given
parser.error('Error: path to test data must be specified. Pass --path to command line')
config_output_filename = options.config_filename
with open(config_output_filename, 'rb') as f_in:
C = pickle.load(f_in)
if C.network == 'resnet50':
import keras_frcnn.resnet as nn
elif C.network == 'vgg':
import keras_frcnn.vgg as nn
# turn off any data augmentation at test time
C.use_horizontal_flips = False
C.use_vertical_flips = False
C.rot_90 = False
img_path = options.test_path
def format_img_size(img, C):
""" formats the image size based on config """
img_min_side = float(C.im_size)
(height,width,_) = img.shape
if width <= height:
ratio = img_min_side/width
new_height = int(ratio * height)
new_width = int(img_min_side)
else:
ratio = img_min_side/height
new_width = int(ratio * width)
new_height = int(img_min_side)
img = cv2.resize(img, (new_width, new_height), interpolation=cv2.INTER_CUBIC)
return img, ratio
def format_img_channels(img, C):
""" formats the image channels based on config """
img = img[:, :, (2, 1, 0)]
img = img.astype(np.float32)
img[:, :, 0] -= C.img_channel_mean[0]
img[:, :, 1] -= C.img_channel_mean[1]
img[:, :, 2] -= C.img_channel_mean[2]
img /= C.img_scaling_factor
img = np.transpose(img, (2, 0, 1))
img = np.expand_dims(img, axis=0)
return img
def format_img(img, C):
""" formats an image for model prediction based on config """
img, ratio = format_img_size(img, C)
img = format_img_channels(img, C)
return img, ratio
# Method to transform the coordinates of the bounding box to its original size
def get_real_coordinates(ratio, x1, y1, x2, y2):
real_x1 = int(round(x1 // ratio))
real_y1 = int(round(y1 // ratio))
real_x2 = int(round(x2 // ratio))
real_y2 = int(round(y2 // ratio))
return (real_x1, real_y1, real_x2 ,real_y2)
class_mapping = C.class_mapping
if 'bg' not in class_mapping:
class_mapping['bg'] = len(class_mapping)
class_mapping = {v: k for k, v in class_mapping.items()}
print(class_mapping)
class_to_color = {class_mapping[v]: np.random.randint(0, 255, 3) for v in class_mapping}
C.num_rois = int(options.num_rois)
if C.network == 'resnet50':
num_features = 1024
elif C.network == 'vgg':
num_features = 512
if K.image_dim_ordering() == 'th':
input_shape_img = (3, None, None)
input_shape_features = (num_features, None, None)
else:
input_shape_img = (None, None, 3)
input_shape_features = (None, None, num_features)
img_input = Input(shape=input_shape_img)
roi_input = Input(shape=(C.num_rois, 4))
feature_map_input = Input(shape=input_shape_features)
# define the base network (resnet here, can be VGG, Inception, etc)
shared_layers = nn.nn_base(img_input, trainable=True)
# define the RPN, built on the base layers
num_anchors = len(C.anchor_box_scales) * len(C.anchor_box_ratios)
rpn_layers = nn.rpn(shared_layers, num_anchors)
classifier = nn.classifier(feature_map_input, roi_input, C.num_rois, nb_classes=len(class_mapping), trainable=True)
model_rpn = Model(img_input, rpn_layers)
model_classifier_only = Model([feature_map_input, roi_input], classifier)
model_classifier = Model([feature_map_input, roi_input], classifier)
print('Loading weights from {}'.format(C.model_path))
model_rpn.load_weights(C.model_path, by_name=True)
model_classifier.load_weights(C.model_path, by_name=True)
model_rpn.compile(optimizer='sgd', loss='mse')
model_classifier.compile(optimizer='sgd', loss='mse')
all_imgs = []
classes = {}
bbox_threshold = 0.8
visualise = True
for idx, img_name in enumerate(sorted(os.listdir(img_path))):
if not img_name.lower().endswith(('.bmp', '.jpeg', '.jpg', '.png', '.tif', '.tiff')):
continue
print(img_name)
st = time.time()
filepath = os.path.join(img_path,img_name)
img = cv2.imread(filepath)
X, ratio = format_img(img, C)
if K.image_dim_ordering() == 'tf':
X = np.transpose(X, (0, 2, 3, 1))
# get the feature maps and output from the RPN
[Y1, Y2, F] = model_rpn.predict(X)
R = roi_helpers.rpn_to_roi(Y1, Y2, C, K.image_dim_ordering(), overlap_thresh=0.7)
# convert from (x1,y1,x2,y2) to (x,y,w,h)
R[:, 2] -= R[:, 0]
R[:, 3] -= R[:, 1]
# apply the spatial pyramid pooling to the proposed regions
bboxes = {}
probs = {}
for jk in range(R.shape[0]//C.num_rois + 1):
ROIs = np.expand_dims(R[C.num_rois*jk:C.num_rois*(jk+1), :], axis=0)
if ROIs.shape[1] == 0:
break
if jk == R.shape[0]//C.num_rois:
#pad R
curr_shape = ROIs.shape
target_shape = (curr_shape[0],C.num_rois,curr_shape[2])
ROIs_padded = np.zeros(target_shape).astype(ROIs.dtype)
ROIs_padded[:, :curr_shape[1], :] = ROIs
ROIs_padded[0, curr_shape[1]:, :] = ROIs[0, 0, :]
ROIs = ROIs_padded
[P_cls, P_regr] = model_classifier_only.predict([F, ROIs])
for ii in range(P_cls.shape[1]):
if np.max(P_cls[0, ii, :]) < bbox_threshold or np.argmax(P_cls[0, ii, :]) == (P_cls.shape[2] - 1):
continue
cls_name = class_mapping[np.argmax(P_cls[0, ii, :])]
if cls_name not in bboxes:
bboxes[cls_name] = []
probs[cls_name] = []
(x, y, w, h) = ROIs[0, ii, :]
cls_num = np.argmax(P_cls[0, ii, :])
try:
(tx, ty, tw, th) = P_regr[0, ii, 4*cls_num:4*(cls_num+1)]
tx /= C.classifier_regr_std[0]
ty /= C.classifier_regr_std[1]
tw /= C.classifier_regr_std[2]
th /= C.classifier_regr_std[3]
x, y, w, h = roi_helpers.apply_regr(x, y, w, h, tx, ty, tw, th)
except:
pass
bboxes[cls_name].append([C.rpn_stride*x, C.rpn_stride*y, C.rpn_stride*(x+w), C.rpn_stride*(y+h)])
probs[cls_name].append(np.max(P_cls[0, ii, :]))
all_dets = []
for key in bboxes:
bbox = np.array(bboxes[key])
new_boxes, new_probs = roi_helpers.non_max_suppression_fast(bbox, np.array(probs[key]), overlap_thresh=0.5)
for jk in range(new_boxes.shape[0]):
(x1, y1, x2, y2) = new_boxes[jk,:]
(real_x1, real_y1, real_x2, real_y2) = get_real_coordinates(ratio, x1, y1, x2, y2)
cv2.rectangle(img,(real_x1, real_y1), (real_x2, real_y2), (int(class_to_color[key][0]), int(class_to_color[key][1]), int(class_to_color[key][2])),2)
textLabel = '{}: {}'.format(key,int(100*new_probs[jk]))
all_dets.append((key,100*new_probs[jk]))
(retval,baseLine) = cv2.getTextSize(textLabel,cv2.FONT_HERSHEY_COMPLEX,1,1)
textOrg = (real_x1, real_y1-0)
cv2.rectangle(img, (textOrg[0] - 5, textOrg[1]+baseLine - 5), (textOrg[0]+retval[0] + 5, textOrg[1]-retval[1] - 5), (0, 0, 0), 2)
cv2.rectangle(img, (textOrg[0] - 5,textOrg[1]+baseLine - 5), (textOrg[0]+retval[0] + 5, textOrg[1]-retval[1] - 5), (255, 255, 255), -1)
cv2.putText(img, textLabel, textOrg, cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 0), 1)
print('Elapsed time = {}'.format(time.time() - st))
print(all_dets)
cv2.imshow('img', img)
cv2.waitKey(0)
# cv2.imwrite('./results_imgs/{}.png'.format(idx),img)
在train_frcnn中，保存了配置信息config.pickle，后续分类网络送入1：1的正负样本
在test_frcnn中，后续的分类网络全部为正样本
纵观整个keras代码，相较tensorflow版本，代码量小了很多，虽然一些代码写的繁琐（看起来有点绕，比如：IOU计算写了2遍，还有一些多余的转换），不过作为keras入手目标检测是相当不错的选择，对一些细节有了更深的把握～