Add files via upload

Author: engineer1109

tools/demo2.py

@@ -0,0 +1,249 @@
+#!/usr/bin/env python
+# -*- coding: UTF-8 -*
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import _init_paths
+import matplotlib
+matplotlib.use('Agg')
+from model.config import cfg
+from model.test import im_detect
+from model.nms_wrapper import nms
+
+from utils.timer import Timer
+import tensorflow as tf
+from matplotlib.font_manager import FontProperties
+zhfont1 = matplotlib.font_manager.FontProperties(fname='/usr/share/fonts/opentype/noto/NotoSansCJK-Bold.ttc')
+import matplotlib.pyplot as plt
+import numpy as np
+import os, cv2
+import argparse
+
+import csv
+import time
+
+from nets.vgg16 import vgg16
+from nets.resnet_v1_rfcn_hole import resnetv1
+import sys
+reload(sys)
+sys.setdefaultencoding('utf8')
+
+
+CLASSES = ('__background__',
+           'dr0', 'dr1', 'dr2', 'dr3')
+
+NETS = {'vgg16': ('vgg16_faster_rcnn_iter_70000.ckpt',),'res101': ('res101_faster_rcnn_iter_200000.ckpt',)}
+DATASETS= {'pascal_voc': ('voc_2007_trainval',),'pascal_voc_0712': ('voc_2007_trainval+voc_2012_trainval',)}
+localtime = time.asctime( time.localtime(time.time()) )
+
+#thresh在demo()设置了CONF_THRESH=0.8只有概率大于0.8才会显示
+#inds被保留的区块序号，区块信息在dets里
+def vis_detections(im, class_name, dets, thresh=0.5 ,image_name='null'):
+    """Draw detected bounding boxes."""
+    inds = np.where(dets[:, -1] >= thresh)[0]
+    if len(inds) == 0:
+        tm=[0,0,2,2]
+        return 0,0,tm,thresh
+    #print(inds)
+    #im = im[:, :, (2, 1, 0)]
+    #fig, ax = plt.subplots(figsize=(12, 12))
+    #ax.imshow(im, aspect='equal')
+    temp=0
+    maxscore = max(dets[:, -1])
+    for i in inds:
+        bbox = dets[i, :4]
+        score = dets[i, -1]
+        if score==maxscore:
+            return class_name,score,bbox,thresh
+        #ax.add_patch(plt.Rectangle((bbox[0], bbox[1]),bbox[2] - bbox[0],bbox[3] - bbox[1], fill=False,edgecolor='red', linewidth=3.5))
+        #ax.text(bbox[0], bbox[1] - 2,'{:s} {:.3f}'.format(class_name, score),bbox=dict(facecolor='blue', alpha=0.5),fontsize=14, color='white')
+
+    #ax.set_title(('{} detection results p({} | box) >= {:.1f}').format(class_name, class_name,thresh),fontsize=14)
+    #plt.axis('off')
+    #plt.tight_layout()
+    #plt.draw()
+    #plt.savefig("/var/www/html/figure/"+image_name)
+
+def vis_detections_onlyone(im, class_name, dets, thresh=0.5):
+    """Draw detected bounding boxes."""
+    inds = np.where(dets[:, -1] >= thresh)[0]
+    if len(inds) == 0:
+        return
+    im = im[:, :, (2, 1, 0)]
+    fig, ax = plt.subplots(figsize=(12, 12))
+    ax.imshow(im, aspect='equal')
+    maxscore = max(dets[:, -1])
+    for i in inds:
+        bbox = dets[i, :4]
+        score = dets[i, -1]
+        if score==maxscore:
+            ax.add_patch(plt.Rectangle((bbox[0], bbox[1]),bbox[2] - bbox[0],bbox[3] - bbox[1], fill=False,edgecolor='red', linewidth=3.5))
+            ax.text(bbox[0], bbox[1] - 2,'{:s} {:.3f}'.format(class_name, score),bbox=dict(facecolor='blue', alpha=0.5),fontsize=14, color='white')
+
+    ax.set_title(('{} detection results '
+                  'p({} | box) >= {:.1f}').format(class_name, class_name,
+                                                  thresh),
+                  fontsize=14)
+    plt.axis('off')
+    plt.tight_layout()
+    plt.draw()
+
+def demo(sess, net, image_name):
+    """Detect object classes in an image using pre-computed object proposals."""
+
+    # Load the demo image
+    im_file = os.path.join(cfg.DATA_DIR, 'demo', image_name)
+    im = cv2.imread(im_file)
+
+    # Detect all object classes and regress object bounds
+    # im_dect at test.py
+    timer = Timer()
+    timer.tic()
+    scores, boxes = im_detect(sess, net, im)
+    #print(scores)
+    #print(boxes)
+    timer.toc()
+    print('检测区域采样时间 {:.3f}s 共计 {:d} 个目标区块'.format(timer.total_time, boxes.shape[0]))
+
+    # Visualize detections for each class
+    CONF_THRESH = 0.5
+    NMS_THRESH = 0.3
+    #enumerate枚举 hstack矩阵拼接 keep记录nms筛选后的区块 dets保存的每个区块的（x1 y1 x2 y2 score)格式list
+    b=0
+    e='error'
+    f=[0,0,2,2]
+    for cls_ind, cls in enumerate(CLASSES[1:]):
+        cls_ind += 1 # because we skipped background
+        cls_boxes = boxes[:, 4*cls_ind:4*(cls_ind + 1)]
+        cls_scores = scores[:, cls_ind]
+        dets = np.hstack((cls_boxes,
+                          cls_scores[:, np.newaxis])).astype(np.float32)
+        keep = nms(dets, NMS_THRESH)
+        #print(keep)
+        dets = dets[keep, :]
+        a=image_name
+        c,d,g,t=vis_detections(im, cls, dets, thresh=CONF_THRESH, image_name=a)
+        if (c==0 and d==0):
+            1+1
+        else:
+            if (d>b):
+                b=d
+                e=c
+                f=g
+    if (e=='dr0'):
+        e='正常人dr0'
+    if (e=='dr1'):
+        e='轻度患者dr1'
+    if (e=='dr2'):
+        e='中度患者dr2'
+    if (e=='dr3'):
+        e='重度患者dr3'
+    if (e=='dr4'):
+        e='增殖患者dr4'
+    print(e,b)
+    im = im[:, :, (2, 1, 0)]
+    fig, ax = plt.subplots(figsize=(8, 6))
+    ax.imshow(im, aspect='equal')
+    ax.add_patch(plt.Rectangle((f[0], f[1]),f[2] - f[0],f[3] - f[1], fill=False,edgecolor='red', linewidth=3.5))
+    ax.text(f[0], f[1] + 25,'{:s} {:.3f}'.format(e, b),bbox=dict(facecolor='blue', alpha=0.5),fontsize=14, color='white',fontproperties=zhfont1)
+
+    ax.set_title(('辅助诊断结果:{} p({} | box) >= {:.1f}').format(e, e,t),fontsize=14,fontproperties=zhfont1)
+    plt.axis('off')
+    plt.tight_layout()
+    plt.draw()
+    #plt.savefig("/var/www/html/figure/"+image_name)
+    plt.savefig("figure/"+localtime+image_name)
+    filer=open('results/result'+localtime,'a+')
+    filer.write(image_name+' '+e+' '+str(b)+'\n')
+    filer.close
+
+def parse_args():
+    """Parse input arguments."""
+    parser = argparse.ArgumentParser(description='Tensorflow demo')
+    parser.add_argument('--net', dest='demo_net', help='Network to use [vgg16 res101]',
+                        choices=NETS.keys(), default='res101')
+    parser.add_argument('--dataset', dest='dataset', help='Trained dataset [pascal_voc pascal_voc_0712]',
+                        choices=DATASETS.keys(), default='pascal_voc')
+    args = parser.parse_args()
+
+    return args
+
+def csv_writer(data, filename):
+    with open(filename, "wb") as csv_file:
+        writer = csv.writer(csv_file)
+        for line in data:
+            writer.writerow(line)
+
+if __name__ == '__main__':
+    cfg.TEST.HAS_RPN = True  # Use RPN for proposals
+    args = parse_args()
+
+    # model path
+    demonet = args.demo_net
+    dataset = args.dataset
+    tfmodel = os.path.join('output', demonet, DATASETS[dataset][0], 'default',
+                              NETS[demonet][0])
+
+
+    if not os.path.isfile(tfmodel + '.meta'):
+        raise IOError(('{:s} not found.\nDid you download the proper networks from '
+                       'our server and place them properly?').format(tfmodel + '.meta'))
+
+    # set config
+    tfconfig = tf.ConfigProto(allow_soft_placement=True)
+    tfconfig.gpu_options.allow_growth=True
+    print ("\033[1;34mThe code is licensed by engineer1109.")
+    #print ("\033[1;34m开始初始化系统")
+    # init session
+    sess = tf.Session(config=tfconfig)
+    #print ("\033[1;33m卷积模型开始加载，默认是RES101")
+    # load network
+    if demonet == 'vgg16':
+        net = vgg16(batch_size=1)
+    elif demonet == 'res101':
+        net = resnetv1(batch_size=1, num_layers=101)
+    else:
+        raise NotImplementedError
+    #print(demonet)
+    net.create_architecture(sess, "TEST", 5,
+                          tag='default', anchor_scales=[8, 16, 32])
+    saver = tf.train.Saver()
+    saver.restore(sess, tfmodel)
+    #print(saver.restore(sess, tfmodel))
+
+    #print('网络加载完毕 {:s}'.format(tfmodel))
+    result=[]
+    fd = file("images.txt", "r" )  
+    for line in fd.readlines():  
+        result.append(list(map(str,line.split(','))))  
+    #print (result)
+    print ("欢迎使用TF-RFCN测试模式</br>")
+    print ("总共需要辨识的图片数量</br>")
+    size=len(result)
+    print (size)
+    image_name = [1]*size
+    for i in range(size):
+        var=str(result[i][0])
+        var=var.strip()
+        image_name[i]=var
+    #print (image_name)
+    #print(type(result))
+    for image_name in image_name:
+        print('</br>=======================</br>')
+        print('====测试-TF-RFCN====</br>')
+        print('测试数据 data/demo/{}</br>'.format(image_name))
+        demo(sess, net, image_name)
+        data = []
+    data2=[]
+    with open('results/result'+localtime) as f:
+        for line in f:
+            data2.append(line.strip().split(" "))
+    #print('</br>概率分布</br>')
+    #print (data2)
+    filename = "csvfiles/output"+localtime+".csv"
+    csv_writer(data2, filename)
+    filename = "output.csv"
+    csv_writer(data2, filename)
+    #plt.show()