Domain adaptation for ear recognition using deep convolutional neural networks

Abstract

Here, the authors have extensively investigated the unconstrained ear recognition problem. The authors have first shown the importance of domain adaptation, when deep convolutional neural network (CNN) models are used for ear recognition. To enable domain adaptation, the authors have collected a new ear data set using the Multi-PIE face data set, which they named as Multi-PIE ear data set. The authors have analysed in depth the effect of ear image quality, for example, illumination and aspect ratio, on the classification performance. Finally, the authors have addressed the problem of data set bias in the ear recognition field. Experiments on the UERC data set have shown that domain adaptation leads to a significant performance improvement. For example, when VGG-16 model is used and the domain adaptation is applied, an absolute increase of around 10% has been achieved. Combining different deep CNN models has further improved the accuracy by 4%. In the experiments that the authors have conducted to examine the data set bias, given an ear image, they were able to classify the data set that it has come from with 99.71% accuracy, which indicates a strong bias among the ear recognition data sets.

Deep Convolutional Neural Networks for Unconstrained Ear Recognition

“…AlexNet [9] is considered a deep CNN architecture compared with previous CNNs such as LeNet-5 [49], and is the winner of the 2012 ImageNet Large Scale Visual Recognition Challenge (ILSVRC-2012) for image classification [50]. As a result, AlexNet has been applied to numerous recognition tasks including ear recognition [39], [51], [52], [42].…”

“…The authors investigated the network's depth on the recognition accuracy using a network depth from 11 to 19 layers. VGGNets have been applied to improve recognition performance on challenging image datasets including the unconstrained ear image datasets [51], [44], [43], [32].…”

“…However, in our experiments we found that an input size of 171 × 235 achieves the best results. While variants of the Inception network (InceptionV1) have been used in various recognition tasks including ear recognition [51], [44], to the best of our knowledge, InceptionV3 has not been utilized in any ear recognition experiments. Due to its superior performance as the first runner-up of the ILSVRC-2015 competition [50], it has been selected for our experiments.…”

Semi-supervised representation learning via dual autoencoders for domain adaptation

“…In recent years, deep learning has been widely used in natural language processing and achieved a satisfying classification performance. Popular deep learning models included Autoencoder [10,11,12,13], Convolutional Neural Network (CNN) [14,15,16], Recurrent Neural Network (RNN) [17,18,19] and Generative Adversarial Network (GAN) [20,21,22,23]. Methods based on autoencoders have been proven to be able to learn domain generic concepts and be beneficial to cross-domain learning tasks.…”

Handcrafted versus CNN Features for Ear Recognition

“…One feasible solution for recognition problems when the amount of data are insufficient is pretraining on a similar recognition task using large scale datasets such as ImageNet [6], as conducted in [7][8][9][10][11]. This technique is referred to as transfer learning and has proven to be effective in plenty of application domains including ear recognition [12][13][14]. In the context of deep CNNs, two types of transfer learning are applicable.…”