在本教程的带领下，你将在十分钟内学会在PYNQ-Z2板卡上使用二值化神经网络开发自己的AI应用！
1. AI初学者入门第一步：分类
AI入门第一步就是对输入的图片进行分类。首先我们需要初始化一个分类器，根据最终实现分类目标的不同需要选择不同的数据集。我们可以在‘road-signs','streeview'和‘cifar10’三个数据集中选择其一，这里选择的是'road-signs'交通标志数据集。
import bnn
print(bnn.available_params(bnn.NETWORK_CNVW1A1))
classifier = bnn.CnvClassifier(bnn.NETWORK_CNVW1A1, 'road-signs', bnn.RUNTIME_HW)
['road-signs', 'streetview', 'cifar10']
2. 一个“训练好”的分类器可以识别多少种目标呢？
我们上面选择的'road-signs'数据集可以识别42种不同的路标标记
print(classifier.classes)
42个Class类别详细请参考阅读原文源代码
3. 接下来我们将用如下几张图片来测试刚刚选择的分类器
from PIL import Image
import numpy as np
from os import listdir
from os.path import isfile, join
from IPython.display import display
imgList = [f for f in listdir("/home/xilinx/jupyter_notebooks/bnn/pictures/road_signs/") if isfile(join("/home/xilinx/jupyter_notebooks/bnn/pictures/road_signs/", f))]
images = []
for imgFile in imgList:
img = Image.open("/home/xilinx/jupyter_notebooks/bnn/pictures/road_signs/" + imgFile)
images.append(img)
img.thumbnail((64, 64), Image.ANTIALIAS)
display(img)
4. 在硬件中对运算进行加速
results = classifier.classify_images(images)
print("Identified classes: {0}".format(results))
for index in results:
print("Identified class name: {0}".format((classifier.class_name(index))))
Inference took 2744.00 microseconds, 914.67 usec per image
Classification rate: 1093.29 images per second
Identified classes: [14 27 41]
Identified class name: Stop
Identified class name: Pedestrians in road ahead
Identified class name: End of no-overtaking zone
5. 作为对比，我们也在软件层面进行一次上面的运算
sw_class = bnn.CnvClassifier(bnn.NETWORK_CNVW1A1,"road-signs", bnn.RUNTIME_SW)
results = sw_class.classify_images(images)
print("Identified classes: {0}".format(results))
for index in results:
print("Identified class name: {0}".format((classifier.class_name(index))))
Inference took 4918608.00 microseconds, 1639536.00 usec per image
Classification rate: 0.61 images per second
Identified classes: [14 27 41]
Identified class name: Stop
Identified class name: Pedestrians in road ahead
Identified class name: End of no-overtaking zone
6. AI初学者入门第二步：识别
如何才能在一大张图片中成功识别到某个具体的物体呢？如下图所示，我们如何从一张十字路口的照片中，识别到中间这个大大的STOP？
from PIL import Image
image_file = "/home/xilinx/jupyter_notebooks/bnn/pictures/street_with_stop.JPG"
im = Image.open(image_file)
im
将一张大图片划分为许多张小图片，并对所有的小图片进行分类预测，就可以得出最有可能是STOP的位置
images = []
bounds = []
for s in [64,96]:
stride = s // 4
x_tiles = im.width // stride
y_tiles = im.height // stride
for j in range(y_tiles):
for i in range(x_tiles):
bound = (stride * i, stride * j, stride * i + s, stride * j + s)
if bound[2] <= im.width and bound[3] < im.height:
c = im.crop(bound)
images.append(c)
bounds.append(bound)
print(len(images))
result = classifier.classify_images_details(images)
result=result.reshape(len(images),43)
from PIL import ImageDraw
draw = ImageDraw.Draw(im)
i=0
for image in images:
# 设定一定的阈值
if result[i][14] > 370:
draw.rectangle(bounds[i], outline='red')
i=i+1
im
1330
Inference took 437603.00 microseconds, 329.02 usec per image
Classification rate: 3039.28 images per second
点击查看源代码：https://github.com/xupsh/pynq-hands-on-demos/tree/master/DeepLearning
文章转载自： Pynq
