前一段时间忙于csgo，也roll了几个小皮肤，沉迷于csgo中，没有学习，有点颓。
今天呢，又跟老师抬杠，抬杠的时候着实很爽，很舒服。害，到头来还是面对现实。
言归正传，我们想要将数字图片喂给神经网络，让神经网络去识别图片上的数字，以验证模型训练的准确性
我们准备了10张数字图片，如下图所示
由前面训练模型，可知我们训练的图片都是28*28像素的黑底白字图片，所以我们得对图片进行预处理
def pre_pic(picName):#对图片进行处理img = Image.open(picName)reIm = img.resize((28,28), Image.ANTIALIAS)#消除图片的一种锯齿差im_arr = np.array(reIm.convert('L'))#把图片变成灰度图，以满足程序的需求threshold = 50for i in range(28):#因为输入的图片是黑底白字，要求的是白底黑字，所以对每一个像素进行反转for j in range(28):im_arr[i][j] = 255 - im_arr[i][j]if (im_arr[i][j] < threshold):im_arr[i][j] = 0else: im_arr[i][j] = 255nm_arr = im_arr.reshape([1, 784])nm_arr = nm_arr.astype(np.float32)img_ready = np.multiply(nm_arr, 1.0/255.0)return img_ready
处理完喂给训练好的神经网络模型
def restore_model(testPicArr):with tf.Graph().as_default() as tg:x = tf.placeholder(tf.float32, [None, mnist_forward.INPUT_NODE])y = mnist_forward.forward(x, None)preValue = tf.argmax(y, 1)variable_averages = tf.train.ExponentialMovingAverage(mnist_backward.MOVING_AVERAGE_DECAY)variables_to_restore = variable_averages.variables_to_restore()saver = tf.train.Saver(variables_to_restore)with tf.Session() as sess:ckpt = tf.train.get_checkpoint_state(mnist_backward.MODEL_SAVE_PATH)if ckpt and ckpt.model_checkpoint_path:saver.restore(sess, ckpt.model_checkpoint_path)preValue = sess.run(preValue, feed_dict={x:testPicArr})return preValueelse:print("No checkpoint file found")return -1
然后输出识别结果，全部代码如下
#coding:utf-8import tensorflow as tfimport numpy as npfrom PIL import Imageimport mnist_backwardimport mnist_forwardimport osos.environ["TF_CPP_MIN_LOG_LEVEL"]='1' # 这是默认的显示等级，显示所有信息 os.environ["TF_CPP_MIN_LOG_LEVEL"]='2' # 只显示 warning 和 Error os.environ["TF_CPP_MIN_LOG_LEVEL"]='3' # 只显示 Errordef restore_model(testPicArr):with tf.Graph().as_default() as tg:x = tf.placeholder(tf.float32, [None, mnist_forward.INPUT_NODE])y = mnist_forward.forward(x, None)preValue = tf.argmax(y, 1)variable_averages = tf.train.ExponentialMovingAverage(mnist_backward.MOVING_AVERAGE_DECAY)variables_to_restore = variable_averages.variables_to_restore()saver = tf.train.Saver(variables_to_restore)with tf.Session() as sess:ckpt = tf.train.get_checkpoint_state(mnist_backward.MODEL_SAVE_PATH)if ckpt and ckpt.model_checkpoint_path:saver.restore(sess, ckpt.model_checkpoint_path)preValue = sess.run(preValue, feed_dict={x:testPicArr})return preValueelse:print("No checkpoint file found")return -1def pre_pic(picName):#对图片进行处理img = Image.open(picName)reIm = img.resize((28,28), Image.ANTIALIAS)#消除图片的一种锯齿差im_arr = np.array(reIm.convert('L'))#把图片变成灰度图，以满足程序的需求threshold = 50for i in range(28):#因为输入的图片是黑底白字，要求的是白底黑字，所以对每一个像素进行反转for j in range(28):im_arr[i][j] = 255 - im_arr[i][j]if (im_arr[i][j] < threshold):im_arr[i][j] = 0else: im_arr[i][j] = 255nm_arr = im_arr.reshape([1, 784])nm_arr = nm_arr.astype(np.float32)img_ready = np.multiply(nm_arr, 1.0/255.0)return img_readydef application():testNum = input("input the number of test pictures:")for i in range(int(testNum)):testPic = input("the path of test picture:")testPicArr = pre_pic(testPic)preValue = restore_model(testPicArr)print( "The prediction number is:", preValue)def main():application()if __name__ == '__main__':main()
为了实验的方便，图片直接以数字大小命名
