def main():
# 【1】搭建网络-->解码网络输出-->设置图片的占位节点
ssd_net = SSD() # 建立SSD模型
classes, scores, bboxes = ssd_net.detections() # 网络中检测函数，返回类别，置信度，回归框位置
images = ssd_net.images() # 给一个图片空间(开辟一个大小)
# 【2】用检查点文件恢复模型
sess = tf.Session()
ckpt_filename = './SSD_model/ssd_vgg_300_weights.ckpt' # 检查点文件
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver()
saver.restore(sess, ckpt_filename) # 参数恢复
# 【3】预处理图片-->处理预测边界框bboxes
img = cv2.imread('./SSD_data/test0.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_prepocessed = preprocess_image(img) # 预处理图片
rclasses, rscores, rbboxes = sess.run([classes, scores, bboxes],
feed_dict={images:
img_prepocessed}) # 检测
rclasses, rscores, rbboxes = process_bboxes(
rclasses, rscores, rbboxes) # 处理预测边界框(剪掉多余的边界，按照置信度保留top-k个框，调整框大小)
# 【4】可视化最终的检测结果
plt_bboxes(img, rclasses, rscores, rbboxes)
print('SSD detection has done!')
"""
import cv2
import numpy as np
import tensorflow as tf
import matplotlib.image as mpimg
from ssd_300_vgg import SSD
from utils import preprocess_image, process_bboxes
from visualization import plt_bboxes
ssd_net = SSD()
classes, scores, bboxes = ssd_net.detections()
images = ssd_net.images()
sess = tf.Session()
# Restore SSD model.
ckpt_filename = './ssd_checkpoints/ssd_vgg_300_weights.ckpt'
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver()
saver.restore(sess, ckpt_filename)
img = cv2.imread('./demo/dog.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_prepocessed = preprocess_image(img)
rclasses, rscores, rbboxes = sess.run([classes, scores, bboxes],
feed_dict={images: img_prepocessed})
rclasses, rscores, rbboxes = process_bboxes(rclasses, rscores, rbboxes)
plt_bboxes(img, rclasses, rscores, rbboxes)
import cv2
import numpy as np
import tensorflow as tf
import matplotlib.image as mpimg
from ssd_300_vgg import SSD
from utils import preprocess_image, process_bboxes
from visualization import plt_bboxes
ssd_net = SSD()
classes, scores, bboxes = ssd_net.detections()
images = ssd_net.images()
sess = tf.Session()
# Restore SSD model.
ckpt_filename = './ssd_checkpoints/ssd_vgg_300_weights.ckpt'
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver()
saver.restore(sess, ckpt_filename)
img = cv2.imread('./demo/dog.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_prepocessed = preprocess_image(img)
rclasses, rscores, rbboxes = sess.run([classes, scores, bboxes],
feed_dict={images: img_prepocessed})
rclasses, rscores, rbboxes = process_bboxes(rclasses, rscores, rbboxes)
plt_bboxes(img, rclasses, rscores, rbboxes)