Densely Connected Convolutional Networks
打开互动全文版（逐段中英对照 + 图/公式 + 论文问答）→近期研究表明，如果卷积网络在靠近输入和输出的层之间包含更短的连接，网络可以变得更深、更准确且更易于训练。本文基于这一观察，提出了密集卷积网络（DenseNet），它以馈送方式将每一层与所有其他层连接。传统的 L 层卷积网络有 L 个连接——每层与其后续层之间一个连接——而我们的网络有 L(L+1)/2 个直接连接。对于每一层，所有前层的特征图被用作输入，其自身的特征图被用作所有后续层的输入。DenseNet 具有几个引人注目的优势：它缓解了梯度消失问题，增强了特征传播，鼓励特征重用，并大幅减少了参数数量。我们在四个极具竞争力的目标识别基准任务（CIFAR-10、CIFAR-100、SVHN 和 ImageNet）上评估了我们提出的架构。DenseNet 在大多数任务上取得了比现有技术显著的改进，同时需要更少的计算量来实现高性能。代码和预训练模型可在 https://github.com/liuzhuang13/DenseNet 获取。
Recent work has shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output. In this paper, we embrace this observation and introduce the Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion. Whereas traditional convolutional networks with L layers have L connections - one between each layer and its subsequent layer - our network has L(L+1)/2 direct connections. For each layer, the feature-maps of all preceding layers are used as inputs, and its own feature-maps are used as inputs into all subsequent layers. DenseNets have several compelling advantages: they alleviate the vanishing-gradient problem, strengthen feature propagation, encourage feature reuse, and substantially reduce the number of parameters. We evaluate our proposed architecture on four highly competitive object recognition benchmark tasks (CIFAR-10, CIFAR-100, SVHN, and ImageNet). DenseNets obtain significant improvements over the state-of-the-art on most of them, whilst requiring less computation to achieve high performance. Code and pre-trained models are available at https://github.com/liuzhuang13/DenseNet .