本文共 1723 字，大约阅读时间需要 5 分钟。
Convolutional networks are powerful visual models that
yield hierarchies of features. 卷积网络是十分有力的在获得层次特征的图像模型当中。We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, exceed the state-of-the-art in semantic segmentation.
我们展示了卷积网络本身，在端到端、像素到像素的语义分割当中，获得了当前最领先的成果。Our key insight is to build “fully convolutional”
我们的主要精力都集中在建立全连接网络。networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning.
网络读入任意大小的输入，并且经过高效的推理和学习输出相应大小的输出。We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models.
我们定义和细化了全卷积网络的空间，并将其的应用作用在空间稠密的预测任务当中，并且和之前的模型相互联系。We adapt contemporary classification networks (AlexNet [20],
the VGG net [31], and GoogLeNet [32]) into fully convolutional networks and transfer their learned representations 我们使得ALexNet、VGG和GooleNet这些当代流行的网络适应于全卷积网络，并且改变他们的学习表现。by fine-tuning [3] to the segmentation task.
通过微调分割任务。 We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. 我们之后定义了一个由语义信息组成的跳跃结构通过一个深度的粗糙的层Our fully convolutional network achieves state-of-the-art segmentation of PASCAL VOC (20% relative improvement to 62.2% mean IU on 2012), NYUDv2, and SIFT
我们的全卷积网络在ALexNet、VGG和GooleNet当中达到了当前阶段最好的的分割效果。 Flow, while inference takes less than one fifth of a second for a typical image.1.搭建了一个全卷积神经网络，输入任意尺寸的图像可以输出相应（我理解为一样）大小的输出。
2.将当前全卷积网络改写成AlexNet和VGGNet和GoogleNet 3.实验结果：PASCAL voc、NYUDv2和SIFT Flow数据集上得到了state-of-the-art的结果，也就是最先进，最好的结果转载地址：http://tkywi.baihongyu.com/