Browsing by Author "Shan, Xiaocai"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item Open Access An adaptive pig face recognition approach using convolutional neural networks(Elsevier, 2020-04-16) Marsot, Mathieu; Mei, Jiangqiang; Shan, Xiaocai; Ye, Liyong; Feng, Peng; Yan, Xuejun; Li, Chenfan; Zhao, YifanThe evolution of agriculture towards intensive farming leads to an increasing demand for animal identification associated with high traceability, driven by the need for quality control and welfare management in agricultural animals. Automatic identification of individual animals is an important step to achieve individualised care in terms of disease detection and control, and improvement of the food quality. For example, as feeding patterns can differ amongst pigs in the same pen, even in homogenous groups, automatic registration shows the most potential when applied to an individual pig. In the EU for instance, this capability is required for certification purposes. Although the RFID technology has been gradually developed and widely applied for this task, chip implanting might still be time-consuming and costly for current practical applications. In this paper, a novel framework composed of computer vision algorithms, machine learning and deep learning techniques is proposed to offer a relatively low-cost and scalable solution of pig recognition. Firstly, pig faces and eyes are detected automatically by two Haar feature-based cascade classifiers and one shallow convolutional neural network to extra high-quality images. Secondly, face recognition is performed by employing a deep convolutional neural network. Additionally, class activation maps generated by grad-CAM and saliency maps are utilised to visually understand how the discriminating parameters have been learned by the neural network. By applying the proposed approach on 10 randomly selected pigs filmed in farm condition, the proposed method demonstrates the superior performance against the state-of-art method with an accuracy of 83% over 320 testing images. The outcome of this study will facilitate the real-application of AI-based animal identification in swine production.Item Open Access Learning spatio-temporal representations with a dual-stream 3-D residual network for nondriving activity recognition(IEEE, 2021-07-28) Yang, Lichao; Shan, Xiaocai; Lv, Chen; Brighton, James; Zhao, YifanAccurate recognition of non-driving activity (NDA) is important for the design of intelligent Human Machine Interface to achieve a smooth and safe control transition in the conditionally automated driving vehicle. However, some characteristics of such activities like limited-extent movement and similar background pose a challenge to the existing 3D convolutional neural network (CNN) based action recognition methods. In this paper, we propose a dual-stream 3D residual network, named D3D ResNet, to enhance the learning of spatio-temporal representation and improve the activity recognition performance. Specifically, a parallel 2-stream structure is introduced to focus on the learning of short-time spatial representation and small-region temporal representation. A 2-feed driver behaviour monitoring framework is further build to classify 4 types of NDAs and 2 types of driving behaviour based on the drivers head and hand movement. A novel NDA dataset has been constructed for the evaluation, where the proposed D3D ResNet achieves 83.35% average accuracy, at least 5% above three selected state-of-the-art methods. Furthermore, this study investigates the spatio-temporal features learned in the hidden layer through the saliency map, which explains the superiority of the proposed model on the selected NDAs.Item Open Access A revised Hilbert-Huang transformation to track non-stationary association of electroencephalography signals(IEEE, 2021-04-28) Shan, Xiaocai; Huo, Shoudong; Yang, Lichao; Cao, Jun; Zou, Jiaru; Chen, Liangyu; Sarrigiannis, Ptolemaios Georgios; Zhao, YifanThe time-varying cross-spectrum method has been used to effectively study transient and dynamic brain functional connectivity between non-stationary electroencephalography (EEG) signals. Wavelet-based cross-spectrum is one of the most widely implemented methods, but it is limited by the spectral leakage caused by the finite length of the basic function that impacts the time and frequency resolutions. This paper proposes a new time-frequency brain functional connectivity analysis framework to track the non-stationary association of two EEG signals based on a Revised Hilbert-Huang Transform (RHHT). The framework can estimate the cross-spectrum of decomposed components of EEG, followed by a surrogate significance test. The results of two simulation examples demonstrate that, within a certain statistical confidence level, the proposed framework outperforms the wavelet-based method in terms of accuracy and time-frequency resolution. A case study on classifying epileptic patients and healthy controls using interictal seizure-free EEG data is also presented. The result suggests that the proposed method has the potential to better differentiate these two groups benefiting from the enhanced measure of dynamic time-frequency association.Item Open Access Spatial–temporal graph convolutional network for Alzheimer classification based on brain functional connectivity imaging of electroencephalogram(Wiley, 2022-06-25) Shan, Xiaocai; Cao, Jun; Huo, Shoudong; Chen, Liangyu; Sarrigiannis, Ptolemaios Georgios; Zhao, YifanFunctional connectivity of the human brain, representing statistical dependence of information flow between cortical regions, significantly contributes to the study of the intrinsic brain network and its functional mechanism. To fully explore its potential in the early diagnosis of Alzheimer's disease (AD) using electroencephalogram (EEG) recordings, this article introduces a novel dynamical spatial–temporal graph convolutional neural network (ST-GCN) for better classification performance. Different from existing studies that are based on either topological brain function characteristics or temporal features of EEG, the proposed ST-GCN considers both the adjacency matrix of functional connectivity from multiple EEG channels and corresponding dynamics of signal EEG channel simultaneously. Different from the traditional graph convolutional neural networks, the proposed ST-GCN makes full use of the constrained spatial topology of functional connectivity and the discriminative dynamic temporal information represented by the 1D convolution. We conducted extensive experiments on the clinical EEG data set of AD patients and Healthy Controls. The results demonstrate that the proposed method achieves better classification performance (92.3%) than the state-of-the-art methods. This approach can not only help diagnose AD but also better understand the effect of normal ageing on brain network characteristics before we can accurately diagnose the condition based on resting-state EEG.Item Open Access Using interictal seizure-free EEG data to recognise patients with epilepsy based on machine learning of brain functional connectivity(Elsevier, 2021-03-12) Cao, Jun; Grajcar, Kacper; Shan, Xiaocai; Zhao, Yifan; Zou, Jiaru; Chen, Liangyu; Li, Zhiqing; Grunewald, Richard; Zis, Panagiotis; De Marco, Matteo; Unwin, Zoe; Blackburn, Daniel; Sarrigiannis, Ptolemaios G.Most seizures in adults with epilepsy occur rather infrequently and as a result, the interictal EEG plays a crucial role in the diagnosis and classification of epilepsy. However, empirical interpretation, of a first EEG in adult patients, has a very low sensitivity ranging between 29-55%. Useful EEG information remains buried within the signals in seizure-free EEG epochs, far beyond the observational capabilities of any specialised physician in this field. Unlike most of the existing works focusing on either seizure data or single-variate method, we introduce a multi-variate method to characterise sensor level brain functional connectivity from interictal EEG data to identify patients with generalised epilepsy. A total of 9 connectivity features based on 5 different measures in time, frequency and time frequency domains have been tested. The solution has been validated by the K-Nearest Neighbour algorithm, classifying an epilepsy group (EG) vs healthy controls (HC) and subsequently with another cohort of patients characterised by non-epileptic attacks (NEAD), a psychogenic type of disorder. A high classification accuracy (97%) was achieved for EG vs HC while revealing significant spatio temporal deficits in the frontocentral areas in the beta frequency band. For EG vs NEAD, the classification accuracy was only about 73%, which might be a reflection of the well-described coexistence of NEAD with epileptic attacks. Our work demonstrates that seizure-free interictal EEG data can be used to accurately classify patients with generalised epilepsy from HC and that more systematic work is required in this direction aiming to produce a clinically useful diagnostic method.