Writer identification has practical applications for forgery detection and
forensic science. Most models based on deep neural networks extract features
from character image or sub-regions in character image, which ignoring features
contained in page-region image. Our proposed global regular network (GRN) pays
attention to these features. GRN network consists of two branches: one branch
takes page handwriting as input to extract global features, and the other takes
word handwriting as input to extract local features. Global features and local
features merge in a global residual way to form overall features of the
handwriting. The proposed GRN has two attributions: one is adding a branch to
extract features contained in page; the other is using residual attention
network to extract local feature. Experiments demonstrate the effectiveness of
both strategies. On CVL dataset, our models achieve impressive 99.98% top-1
accuracy and 100% top-5 accuracy with shorter training time and fewer network
parameters, which exceeded the state-of-the-art structure. The experiment shows
the powerful ability of the network in the field of writer identification. The
source code is available at this https URL.

提出了一种全局正则网络（GRN），其由一个带有全局特征提取器的分支和一个带有局部特征提取器的分支组成，两者以一种全局残差方式融合，可以有效地识别手写字的写作人，实验表明该模型在 CVL 数据集上取得了卓越的 99.98％-1 准确率和 100％ top-5 准确率，具有更短的训练时间和更少的网络参数。

全球正则网络用于笔迹识别

Global Regular Network for Writer Identification

In this work, we propose "Residual Attention Network", a convolutional neural
network using attention mechanism which can incorporate with state-of-art feed
forward network architecture in an end-to-end training fashion. Our Residual
Attention Network is built by stacking Attention Modules which generate
attention-aware features. The attention-aware features from different modules
change adaptively as layers going deeper. Inside each Attention Module,
bottom-up top-down feedforward structure is used to unfold the feedforward and
feedback attention process into a single feedforward process. Importantly, we
propose attention residual learning to train very deep Residual Attention
Networks which can be easily scaled up to hundreds of layers. Extensive
analyses are conducted on CIFAR-10 and CIFAR-100 datasets to verify the
effectiveness of every module mentioned above. Our Residual Attention Network
achieves state-of-the-art object recognition performance on three benchmark
datasets including CIFAR-10 (3.90% error), CIFAR-100 (20.45% error) and
ImageNet (4.8% single model and single crop, top-5 error). Note that, our
method achieves 0.6% top-1 accuracy improvement with 46% trunk depth and 69%
forward FLOPs comparing to ResNet-200. The experiment also demonstrates that
our network is robust against noisy labels.

通过引入注意力机制，在端到端的训练过程中将其与最先进的前馈网络结构相结合，我们提出了一种卷积神经网络模型 “Residual Attention Network”，并在 CIFAR-10、CIFAR-100 和 ImageNet 三个基准数据集上实现了最佳的物体识别表现。