Analyzing features learned for Offline Signature Verification using

Analyzing features learned for Offline Signature Verification using Deep CNNs 离线签名鉴别使用Deep CNNs进行分析特征学习 报告人：赵妍 汇报时间： 20161031

本文方法 l 中心思想：Learn a features representation(a function Φ(. ) for offline signature verification in a Writer-Independent format , use this function to extract features from signatures X of the users enrolled in the system(Φ(. ) ), and use the resulting feature vectors to train a binary classifier for each user. ) l 中心思想就是在书写者独立的情况下为离线签名鉴别学习一种 特征表示方法（一个函数Φ（. ）），使用这个函数去从系统（ Φ（X））中登记了的签名X中提取特征，然后使用结果特征向 量为每一个用户训练一个二值分类器。 Slide No. 4

本文方法 l 在书写者独立的情况下学习特征表示有两个基本原理： ① 直接为每个用户学习特征表示是不切实际的，提供大约 5 -10个 签名的小样本样品用来训练； ② 通过使用模型来为新的签名者的真实签名提取特征，使得可以 将新的用户加入系统，训练一个写作者依赖的分类器。（having a fixed representation useful for any user makes it straightforward to add new users to the system, by simply using the model to extract features for the new user’s genuine signatures, and training a Writer. Dependent classifier） Slide No. 5

本文方法 Overall, the method consists in the following steps: 1. Training a deep neural network on a Development set 2. Using this network to obtain a new representation for signatures on E (i. e. obtain Φ(X) for all signatures X) 3. Training Writer-Dependent classifiers in the exploitation set, using the learned representation Slide No. 7

本文方法---CNN训练 l As in [9], we learn a function Φ(. ) by training a Deep Convolutional Neural Network on a Development set, by learning to discriminate between different users. l 如文章 9所说，我们学习一个函数Φ（. ）通过在一个发展集上训练 一个深的卷积神经网络，通过学习不同用户间的区别。 l That is, we model the network to output M units, that estimate P(y|X) where y is one of the M users in the set D, and X is a signature image. l 也就是，我们模拟网络输出M单元，估计P（y|X），其中y是集合D 中M使用者中的一个，X是一个签名图像。 Slide No. 8

本文方法---CNN训练 l In this work we investigate different architectures for learning feature representations. In particular, we evaluate the impact of depth, and the impact of the size of the embedding layer (the layer from which we obtain the representation of the signature). l 在这一 作中，我们为学习特征表示尝试不同的架构。尤其是，我 们评价其深度的影响，以及嵌入层大小的影响（我们从中获得签名 的表示的层）。 Slide No. 9

本文方法---训练WD分类器 l 在发展集D上训练了Deep CNN之后，使用该网络从另一个用户集合E 中提取签名的特征表示，训练WD（写作者依赖）的分类器。 l 做法： we resize the images to 170 x 242 pixels, perform feedforward propagation until a fully-connect layer, and used the activations at that layer as the feature vector for the image. For the architectures marked as “reduced” (see table I), that contained only two fully-connected layers, we consider only the last layer before softmax as the embedding layer (FC 1). For the architectures with three fully-connected layers, we consider both layers FC 1 and FC 2, that is, the last two fully-connected layers before the softmax layer. Slide No. 13

谢谢！ Slide No. 24