tw记账

CMU Multi-PIE人脸数据库介绍
http://yongfeng.me/
唇语识别 
https://github.com/hzy46/fast-neural-style-tensorflow
动力外骨骼机器人
python tensorflow  生成对抗网络 差异比较
Image Inpainting
梯形 矫正
立体定向
Semantic Image Segmentation via Deep Parsing Network
Perceptual Losses
okay, i found the solution to solve it， i tried to python3 setup.py build and install in research folder, after done that, i executed the code,and worked
先下载你要安装的包，并解压到磁盘下；
进入到该文件的setup.py 目录下 ，打开cmd，并切换到该目录下；
先执行 python setup.py build
然后执行 python setup.py install
pip install --upgrade https://mirrors.tuna.tsinghua.edu.cn/tensorflow/windows/gpu/tensorflow_gpu-1.3.0rc0-cp35-cp35m-win_amd64.whl


exit(0)
MIAS标准数据集
Windows Driver Kit
Image Completion 
aiimooc
1 
https://baijia.baidu.com/s?id=1575151965430656
[e-mail]
http://www.lfd.uci.edu/~gohlke/pythonlibs/#scikit-image
ceil
https://orcid.org/
0000-0003-2698-8507下载数据库
 DDSM
下载数据库  转换 泡一个模型 后改进 。3个月
traffic sign detection
https://luna16.grand-challenge.org/

https://www.kaggle.com/arnavkj95/candidate-generation-and-luna16-preprocessing
Image tamper detection
“贴心小棉袄” iTBra智能文胸检测女性乳房癌症
pip install --upgrade
 这个是乳腺癌的，RIDER Breast 乳腺癌MRI影像数据，400M左右，http://dataju.cn/Dataju/web/datasetInstanceDetail/415 
点配准算法
75 16
# 最小化方差loss = tf.reduce_mean(tf.square(y - y_data))
0615  目标之后添加 高斯 噪声 ？？或许 测试 均匀噪声 ？？
super resolution
def load_model(session,netmodel_path,param_path):   new_saver = tf.train.import_meta_graph(netmodel_path)   new_saver.restore(session, param_path)   x= tf.get_collection('real_A')[0]#在训练阶段需要调用tf.add_to_collection('test_images',test_images),保存之   y = tf.get_collection("fake_B")[0]   return  x,y

y_np=sess.run(tensorB,feed_dict = {tensorA:images})

zxing  zba

码湿度，解码 ，二维码以外

神经外科导航定位机器人

https://gist.github.com/33c758ad77e6e6531392

SimGAN.

https://gist.github.com/33c758ad77e6e6531392

output_list = []for j in range(0,16): show_img1 = features[j, :, :, :] #lab = show_img1.reshape([1] + list(show_img1.shape)).astype(np.float32) #show_img2 = sess.run(show_img1) #show_img2 = tf.reshape(show_img1, [64, 64, 3]) show_img2 = tf.cast(show_img1, tf.float32)/255.0 #show_img = tf.image.resize_images(show_img, [16, 16]) #show_img3 = show_img2 show_img3 = salt(show_img2, 7) show_img4 = show_img3 show_img5 = ops.convert_to_tensor(show_img4, dtype=tf.float32,name='labels_and_features') #show_img3 = tf.reshape(show_img2, [64, 64, 3]) #lab = show_img.reshape([1] + list(show_img.shape)).astype(np.float32) #jab = show_img3 #train_features[i, :, :, :] = lab #lab = tf.reshape(lab, [64, 64, 3]) output_list.append(show_img5) #lab2 = tf.reshape(output_list, [16, 64, 64, 3]) #lab = tf.reshape(lab, [16,64, 64, 3]) #lab=tf.convert_to_tensor(output_list)jab2 = ops.convert_to_tensor(output_list, dtype=tf.float32, name='labels_and_features')

train_features2 = train_featurestest_features2 = test_featurestrain_features1 = sess.run(train_features)test_features1 = sess.run(test_features)
php  imagefilter  imagecolorstotal
yum install python-imaging

随机过程论（三节连上）｛1-16周[教师:孙洪祥,地点:3-235]｝星期四 前三节

import cv2import numpy as npdef salt(img, img2,n): noff=int((n-1)/2)   for i in range(img.shape[1]):       for j in range(img.shape[1]):        # img.shape[0] -- 取得img 的列（图片的高）        # img.shape[1] -- 取得img 的行（图片的宽）        #i = int(np.random.random() * img.shape[1]);        #j = int(np.random.random() * img.shape[0]);                   img[j, i, 0] = 255            img[j, i, 1] = 255            img[j, i, 2] = 255       return imgimg = cv2.imread("000001.jpg")img2 = img.copy()saltImage = salt(img, img2,11)cv2.imshow("Salt", saltImage)cv2.waitKey(0)cv2.destroyAllWindows()

http://blog.csdn.net/chinamming?viewmode=contents

https://public.kitware.com/IGSTKWIKI/index.php/How_to_Build_IGSTK(old)