The use of deep learning methods to extract vascular biometric patterns from the palm surface has been of interest among researchers in recent years. In many biometric recognition tasks, there is a limit in the number of training samples. This is because of limited vein biometric databases being available for research. This restricts the application of deep learning methods to design algorithms that can effectively identify or authenticate people for vein recognition. This paper proposes an architecture using Siamese neural network structure for few shot palm vein identification. The proposed network uses images from both the palms and consists of two sub-nets that share weights to identify a person. The architecture performance was tested on the HK PolyU multi spectral palm vein database with limited samples. The results suggest that the method is effective since it has 91.9% precision, 91.1% recall, 92.2% specificity, 91.5%, F1-Score, and 90.5% accuracy values.
With the increase in computation power and the development of new state-of-the-art deep learning algorithms, appearance-based gaze estimation is becoming more and more popular. It is believed to work well with curated laboratory data sets, however it faces several challenges when deployed in real world scenario. One such challenge is to estimate the gaze of a person about which the Deep Learning model trained for gaze estimation has no knowledge about. To analyse the performance in such scenarios we have tried to simulate a calibration mechanism. In this work we use the MPIIGaze data set. We trained a multi modal convolutional neural network and analysed its performance with and without calibration and this evaluation provides clear insights on how calibration improved the performance of the Deep Learning model in estimating gaze in the wild.
Deep generative networks in recent years have reinforced the need for caution while consuming various modalities of digital information. One avenue of deepfake creation is aligned with injection and removal of tumors from medical scans. Failure to detect medical deepfakes can lead to large setbacks on hospital resources or even loss of life. This paper attempts to address the detection of such attacks with a structured case study. Specifically, we evaluate eight different machine learning algorithms, which including three conventional machine learning methods, support vector machine, random forest, decision tree, and five deep learning models, DenseNet121, DenseNet201, ResNet50, ResNet101, VGG19, on distinguishing between tampered and untampered images.For deep learning models, the five models are used for feature extraction, then fine-tune for each pre-trained model is performed. The findings of this work show near perfect accuracy in detecting instances of tumor injections and removals.
Circulating Tumor Cells (CTCs) bear great promise as biomarkers in tumor prognosis. However, the process of identification and later enumeration of CTCs require manual labor, which is error-prone and time-consuming. The recent developments in object detection via Deep Learning using Mask-RCNNs and wider availability of pre-trained models have enabled sensitive tasks with limited data of such to be tackled with unprecedented accuracy. In this report, we present a novel 3-stage detection model for automated identification of Circulating Tumor Cells in multi-channel darkfield microscopic images comprised of: RetinaNet based identification of Cytokeratin (CK) stains, Mask-RCNN based cell detection of DAPI cell nuclei and Otsu thresholding to detect CD-45s. The training dataset is composed of 46 high variance data points, with 10 Negative and 36 Positive data points. The test set is composed of 420 negative data points. The final accuracy of the pipeline is 98.81%.
Deterministic approaches using iterative optimisation have been historically successful in diffeomorphic image registration (DiffIR). Although these approaches are highly accurate, they typically carry a significant computational burden. Recent developments in stochastic approaches based on deep learning have achieved sub-second runtimes for DiffIR with competitive registration accuracy, offering a fast alternative to conventional iterative methods. In this paper, we attempt to reduce this difference in speed whilst retaining the performance advantage of iterative approaches in DiffIR. We first propose a simple iterative scheme that functionally composes intermediate non-stationary velocity fields to handle large deformations in images whilst guaranteeing diffeomorphisms in the resultant deformation. We then propose a convex optimisation model that uses a regularisation term of arbitrary order to impose smoothness on these velocity fields and solve this model with a fast algorithm that combines Nesterov gradient descent and the alternating direction method of multipliers (ADMM). Finally, we leverage the computational power of GPU to implement this accelerated ADMM solver on a 3D cardiac MRI dataset, further reducing runtime to less than 2 seconds. In addition to producing strictly diffeomorphic deformations, our methods outperform both state-of-the-art deep learning-based and iterative DiffIR approaches in terms of dice and Hausdorff scores, with speed approaching the inference time of deep learning-based methods.
Long Short-Term Memory (LSTM) neural networks have been widely used for time series forecasting problems. However, LSTMs are prone to overfitting and performance reduction during test phases. Several different regularization techniques have been shown in literature to prevent overfitting problems in neural networks. In this paper, first, we introduce application of kernel flow methods for time series forecasting in general. Afterward, we examine the effectiveness of applying kernel flow regularization on LSTM layers to avoid overfitting problems. We describe a regularization method by applying kernel flow loss function on LSTM layers. In experimental results, we show that kernel flow outperforms baseline models on time series forecasting benchmarks. We also compare the effect of dropout and kernel flow regularization techniques on LSTMs. The experimental results illustrate that kernel flow achieves similar regularization effect to dropout. It also shows that the best results is obtained using both kernel flow and dropout regularizations with early stopping on LSTM layers on some time series datasets (e.g. power-load demand forecasts).
COVID-19 has driven most schools to remote learning through online meeting software such as Zoom and Google Meet. Although this trend helps students continue learning without in-person classes, it removes a vital tool that teachers use to teach effectively: visual cues. By not being able to see a student's face clearly, the teacher may not notice when the student needs assistance, or when the student is not paying attention. In order to help remedy the teachers of this challenge, this project proposes a machine learning based approach that provides real-time student mental state monitoring and classifications for the teachers to better conduct remote teaching. Using publicly available electroencephalogram (EEG) data collections, this research explored four different classification techniques: the classic deep neural network, the traditionally popular support vector machine, the latest convolutional neural network, and the XGBoost model, which has gained popularity recently. This study defined three mental classes: an engaged learning mode, a confused learning mode, and a relaxed mode. The experimental results from this project showed that these selected classifiers have varying potentials in classifying EEG signals for mental states. While some of the selected classifiers only yield around 50% accuracy with some delay, the best ones can achieve 80% accurate classification in real-time. This could be very beneficial for teachers in need of help making remote teaching adjustments, and for many other potential applications where in-person interactions are not possible.
Deep learning techniques have emerged as a promising approach to highly accelerated MRI. However, recent reconstruction challenges have shown several drawbacks in current deep learning approaches, including the loss of fine image details even using models that perform well in terms of global quality metrics. In this study, we propose an end-to-end deep learning framework for image reconstruction and pathology detection, which enables a clinically aware evaluation of deep learning reconstruction quality. The solution is demonstrated for a use case in detecting meniscal tears on knee MRI studies, ultimately finding a loss of fine image details with common reconstruction methods expressed as a reduced ability to detect important pathology like meniscal tears. Despite the common practice of quantitative reconstruction methodology evaluation with metrics such as SSIM, impaired pathology detection as an automated pathology-based reconstruction evaluation approach suggests existing quantitative methods do not capture clinically important reconstruction outcomes.
Diagnostic imaging plays a critical role in healthcare, serving as a fundamental asset for timely diagnosis, disease staging and management as well as for treatment choice, planning, guidance, and follow-up. Among the diagnostic imaging options, ultrasound imaging is uniquely positioned, being a highly cost-effective modality that offers the clinician an unmatched and invaluable level of interaction, enabled by its real-time nature. Ultrasound probes are becoming increasingly compact and portable, with the market demand for low-cost pocket-sized and (in-body) miniaturized devices expanding. At the same time, there is a strong trend towards 3D imaging and the use of high-frame-rate imaging schemes; both accompanied by dramatically increasing data rates that pose a heavy burden on the probe-system communication and subsequent image reconstruction algorithms. With the demand for high-quality image reconstruction and signal extraction from less (e.g unfocused or parallel) transmissions that facilitate fast imaging, and a push towards compact probes, modern ultrasound imaging leans heavily on innovations in powerful digital receive channel processing. Beamforming, the process of mapping received ultrasound echoes to the spatial image domain, naturally lies at the heart of the ultrasound image formation chain. In this chapter on Deep Learning for Ultrasound Beamforming, we discuss why and when deep learning methods can play a compelling role in the digital beamforming pipeline, and then show how these data-driven systems can be leveraged for improved ultrasound image reconstruction.
Deep learning methods have proven to outperform traditional computer vision methods in various areas of image processing. However, the application of deep learning in industrial surface defect detection systems is challenging due to the insufficient amount of training data, the expensive data generation process, the small size, and the rare occurrence of surface defects. From literature and a polymer products manufacturing use case, we identify design requirements which reflect the aforementioned challenges. Addressing these, we conceptualize design principles and features informed by deep learning research. Finally, we instantiate and evaluate the gained design knowledge in the form of actionable guidelines and strategies based on an industrial surface defect detection use case. This article, therefore, contributes to academia as well as practice by (1) systematically identifying challenges for the industrial application of deep learning-based surface defect detection, (2) strategies to overcome these, and (3) an experimental case study assessing the strategies' applicability and usefulness.
In this paper, we propose a deep learning model for Demodulation Reference Signal (DMRS) based channel estimation task. Specifically, a novel Denoise, Linear interpolation and Refine (DLR) pipeline is proposed to mitigate the noise propagation problem during channel information interpolation and to restore the nonlinear variation of wireless channel over time. At the same time, the Small-norm Sample Cost-sensitive (SSC) learning method is proposed to equalize the qualities of channel estimation under different kinds of wireless environments and improve the channel estimation reliability. The effectiveness of the propose DLR-SSC model is verified on WAIC Dataset. Compared with the well know ChannelNet channel estimation model, our DLR-SSC model reduced normalized mean square error (NMSE) by 27.2dB, 22.4dB and 16.8dB respectively at 0dB, 10dB, and 20dB SNR. The proposed model has won the second place in the 2nd Wireless Communication Artificial Intelligence Competition (WAIC). The code is about to open source.
Structures in the solar corona are the main drivers of space weather processes that might directly or indirectly affect the Earth. Thanks to the most recent space-based solar observatories, with capabilities to acquire high-resolution images continuously, the structures in the solar corona can be monitored over the years with a time resolution of minutes. For this purpose, we have developed a method for automatic segmentation of solar corona structures observed in EUV spectrum that is based on a deep learning approach utilizing Convolutional Neural Networks. The available input datasets have been examined together with our own dataset based on the manual annotation of the target structures. Indeed, the input dataset is the main limitation of the developed model's performance. Our \textit{SCSS-Net} model provides results for coronal holes and active regions that could be compared with other generally used methods for automatic segmentation. Even more, it provides a universal procedure to identify structures in the solar corona with the help of the transfer learning technique. The outputs of the model can be then used for further statistical studies of connections between solar activity and the influence of space weather on Earth.
A framework was developed to address limitations associated with existing techniques for analysing sequences. This work deals with the steps followed to select suitable datasets characterised by discrete irregular sequential patterns. To identify, select, explore and evaluate which datasets from various sources extracted from more than 400 research articles, an interquartile range method for outlier calculation and a qualitative Billauer's algorithm was adapted to provide periodical peak detection in such datasets. The developed framework was then tested using the most appropriate datasets. The research concluded that the financial market-daily currency exchange domain is the most suitable kind of data set for the evaluation of the designed deep learning framework, as it provides high levels of discrete irregular patterns.
Investors make investment decisions depending on several factors such as fundamental analysis, technical analysis, and quantitative analysis. Another factor on which investors can make investment decisions is through sentiment analysis of news headlines, the sole purpose of this study. Natural Language Processing techniques are typically used to deal with such a large amount of data and get valuable information out of it. NLP algorithms convert raw text into numerical representations that machines can easily understand and interpret. This conversion can be done using various embedding techniques. In this research, embedding techniques used are BoW, TF-IDF, Word2Vec, BERT, GloVe, and FastText, and then fed to deep learning models such as RNN and LSTM. This work aims to evaluate these model's performance to choose the robust model in identifying the significant factors influencing the prediction. During this research, it was expected that Deep Leaming would be applied to get the desired results or achieve better accuracy than the state-of-the-art. The models are compared to check their outputs to know which one has performed better.
With the recent increase in the computational power of modern mobile devices, machine learning-based heavy tasks such as face detection and speech recognition are now integral parts of such devices. This requires frameworks to execute machine learning models (e.g., Deep Neural Networks) on mobile devices. Although there exist studies on the accuracy and performance of these frameworks, the quality of on-device deep learning frameworks, in terms of their robustness, has not been systematically studied yet. In this paper, we empirically compare two on-device deep learning frameworks with three adversarial attacks on three different model architectures. We also use both the quantized and unquantized variants for each architecture. The results show that, in general, neither of the deep learning frameworks is better than the other in terms of robustness, and there is not a significant difference between the PC and mobile frameworks either. However, in cases like Boundary attack, mobile version is more robust than PC. In addition, quantization improves robustness in all cases when moving from PC to mobile.
Deep learning-based image retrieval has been emphasized in computer vision. Representation embedding extracted by deep neural networks (DNNs) not only aims at containing semantic information of the image, but also can manage large-scale image retrieval tasks. In this work, we propose a deep learning-based image retrieval approach using homography transformation augmented contrastive learning to perform large-scale synthetic aperture radar (SAR) image search tasks. Moreover, we propose a training method for the DNNs induced by contrastive learning that does not require any labeling procedure. This may enable tractability of large-scale datasets with relative ease. Finally, we verify the performance of the proposed method by conducting experiments on the polarimetric SAR image datasets.
Deep learning has transformed the way we think of software and what it can do. But deep neural networks are fragile and their behaviors are often surprising. In many settings, we need to provide formal guarantees on the safety, security, correctness, or robustness of neural networks. This book covers foundational ideas from formal verification and their adaptation to reasoning about neural networks and deep learning.
Natural gradient descent (NGD) provided deep insights and powerful tools to deep neural networks. However the computation of Fisher information matrix becomes more and more difficult as the network structure turns large and complex. This paper proposes a new optimization method whose main idea is to accurately replace the natural gradient optimization by reconstructing the network. More specifically, we reconstruct the structure of the deep neural network, and optimize the new network using traditional gradient descent (GD). The reconstructed network achieves the effect of the optimization way with natural gradient descent. Experimental results show that our optimization method can accelerate the convergence of deep network models and achieve better performance than GD while sharing its computational simplicity.
Myocardial T1 mapping is a cardiac MRI technique, used to assess myocardial fibrosis. In this technique, a series of T1-weighted MRI images are acquired with different saturation or inversion times. These images are fitted to the T1 model to estimate the model parameters and construct the desired T1 maps. In the presence of motion, the different T1-weighted images are not aligned. This, in turn, will cause errors and inaccuracies in the final estimation of the T1 maps. Therefore, motion correction is a necessary process for myocardial T1 mapping. We present a deep-learning (DL) based system for cardiac T1-weighted MRI images motion correction. When applying our DL-based motion correction system we achieve a statistically significant improved performance by means of R2 of the model fitting regression, in compared to the model fitting regression without motion correction (0.52 vs 0.29, p<0.05).
Rolling bearings are subject to various faults due to its long-time operation under harsh environment, which will lead to unexpected breakdown of machinery system and cause severe accidents. Deep learning methods recently have gained growing interests and extensively applied in the data-driven bearing fault diagnosis. However, current deep learning methods perform the bearing fault diagnosis in the form of deterministic classification, which overlook the uncertainties that inevitably exist in actual practice. To tackle this issue, in this research we develop a probabilistic fault diagnosis framework that can account for the uncertainty effect in prediction, which bears practical significance. This framework fully leverages the probabilistic feature of Gaussian process classifier (GPC). To facilitate the establishment of high-fidelity GPC, the tailored feature extraction with dimensionality reduction method can be optimally determined through the cross validation-based grid search upon a prespecified method pool consisting of various kernel principal component analysis (KPCA) methods and stacked autoencoder. This strategy can ensure the complex nonlinear relations between the features and faults to be adequately characterized. Furthermore, the sensor fusion concept is adopted to enhance the diagnosis performance. As compared with the traditional deep learning methods, this proposed framework usually requires less labeled data and less effort for parameter tuning. Systematic case studies using the publicly accessible experimental rolling bearing dataset are carried out to validate this new framework. Various influencing factors on fault diagnosis performance also are thoroughly investigated.