Multi-stream Convolutional Neural Networks Research Articles

Mild cognitive impairment prediction based on multi-stream convolutional neural networks

BackgroundMild cognitive impairment (MCI) is the transition stage between the cognitive decline expected in normal aging and more severe cognitive decline such as dementia. The early diagnosis of MCI plays an important role in human healthcare. Current methods of MCI detection include cognitive tests to screen for executive function impairments, possibly followed by neuroimaging tests. However, these methods are expensive and time-consuming. Several studies have demonstrated that MCI and dementia can be detected by machine learning technologies from different modality data. This study proposes a multi-stream convolutional neural network (MCNN) model to predict MCI from face videos.ResultsThe total effective data are 48 facial videos from 45 participants, including 35 videos from normal cognitive participants and 13 videos from MCI participants. The videos are divided into several segments. Then, the MCNN captures the latent facial spatial features and facial dynamic features of each segment and classifies the segment as MCI or normal. Finally, the aggregation stage produces the final detection results of the input video. We evaluate 27 MCNN model combinations including three ResNet architectures, three optimizers, and three activation functions. The experimental results showed that the ResNet-50 backbone with Swish activation function and Ranger optimizer produces the best results with an F1-score of 89% at the segment level. However, the ResNet-18 backbone with Swish and Ranger achieves the F1-score of 100% at the participant level.ConclusionsThis study presents an efficient new method for predicting MCI from facial videos. Studies have shown that MCI can be detected from facial videos, and facial data can be used as a biomarker for MCI. This approach is very promising for developing accurate models for screening MCI through facial data. It demonstrates that automated, non-invasive, and inexpensive MCI screening methods are feasible and do not require highly subjective paper-and-pencil questionnaires. Evaluation of 27 model combinations also found that ResNet-50 with Swish is more stable for different optimizers. Such results provide directions for hyperparameter tuning to further improve MCI predictions.

Cascade Windows-Based Multi-Stream Convolutional Neural Networks Framework for Early Detecting In-Sleep Stroke Using Wristbands

Cascade Windows-Based Multi-Stream Convolutional Neural Networks Framework for Early Detecting In-Sleep Stroke Using Wristbands

Human Fall Detection Using 3D Multi-Stream Convolutional Neural Networks with Fusion.

Human falls, especially for elderly people, can cause serious injuries that might lead to permanent disability. Approximately 20-30% of the aged people in the United States who experienced fall accidents suffer from head trauma, injuries, or bruises. Fall detection is becoming an important public healthcare problem. Timely and accurate fall incident detection could enable the instant delivery of medical services to the injured. New advances in vision-based technologies, including deep learning, have shown significant results in action recognition, where some focus on the detection of fall actions. In this paper, we propose an automatic human fall detection system using multi-stream convolutional neural networks with fusion. The system is based on a multi-level image-fusion approach of every 16 frames of an input video to highlight movement differences within this range. This results of four consecutive preprocessed images are fed to a new proposed and efficient lightweight multi-stream CNN model that is based on a four-branch architecture (4S-3DCNN) that classifies whether there is an incident of a human fall. The evaluation included the use of more than 6392 generated sequences from the Le2i fall detection dataset, which is a publicly available fall video dataset. The proposed method, using three-fold cross-validation to validate generalization and susceptibility to overfitting, achieved a 99.03%, 99.00%, 99.68%, and 99.00% accuracy, sensitivity, specificity, and precision, respectively. The experimental results prove that the proposed model outperforms state-of-the-art models, including GoogleNet, SqueezeNet, ResNet18, and DarkNet19, for fall incident detection.

Multi-Stream Fast Fourier Convolutional Neural Network for Automatic Target Recognition of Ground Military Vehicle

The synthetic aperture radar (SAR) is very useful in both military and civilian applications due to its 24/7, all-weather, and high resolution capabilities, as well as its ability to recognize camouflage and penetrating cover. In the field of SAR image interpretation, target recognition is an important research challenge for researchers all over the world. With the application of high-resolution SAR, the imaging area has been expanding, and different imaging modes have appeared one after another. There are many difficulties with the conventional understanding of human interpretation. There are issues like slow movement, a lot of labor, and poor judgment. Technology for intelligent interpretation needs to be developed immediately. Although deep convolutional neural networks (CNNs) have proven extremely efficient in image recognition, one of the major drawbacks is that they require more parameters as their layers increase. The cost of convolution operation for all convolutional layers is therefore high, and learning lag results from the inevitable rise in computation as the size of the image kernel grows. This study proposes a three ways input of SAR images into multi-stream fast Fourier convolutional neural network (MS-FFCNN). The technique elaborates on the transformation of rudimentary multi-stream CNN into MS-FFCNN. By utilizing the fast Fourier transformation instead of the standard convolution, it lowers the cost of image convolution in CNNs, which lowers the overall computational cost. The multiple streams of FFCNN overcome the problem of insufficient sample size and further improve the long training time while also improving the recognition accuracy. The proposed method yielded good recognition accuracy of 99.92%.

Multi-Stream Convolutional Neural Networks for Rotating Machinery Fault Diagnosis under Noise and Trend Items.

In recent years, rotating machinery fault diagnosis methods based on convolutional neural network have achieved much success. However, in real industrial environments, interfering signals are unavoidable, which may reduce the accuracy of fault diagnosis seriously. Most of the current fault diagnosis methods are of single input type, which may lead to the information contained in the vibration signal not being fully utilized. In this study, theoretical analysis and comprehensive comparative experiments are completed to investigate the time domain input, frequency domain input, and two types of time–frequency domain input. Based on this, a new fault diagnosis model, named multi-stream convolutional neural network, is developed. The model takes the time domain, frequency domain, and time–frequency domain images as input, and it automatically fuses the information contained in different inputs. The proposed model is tested based on three public datasets. The experimental results suggested that the model achieved pretty high accuracy under noise and trend items without the help of signal separation algorithms. In addition, the positive implications of multiple inputs and information fusion are analyzed through the visualization of learned features.

An evolving ensemble model of multi-stream convolutional neural networks for human action recognition in still images

Still image human action recognition (HAR) is a challenging problem owing to limited sources of information and large intra-class and small inter-class variations which requires highly discriminative features. Transfer learning offers the necessary capabilities in producing such features by preserving prior knowledge while learning new representations. However, optimally identifying dynamic numbers of re-trainable layers in the transfer learning process poses a challenge. In this study, we aim to automate the process of optimal configuration identification. Specifically, we propose a novel particle swarm optimisation (PSO) variant, denoted as EnvPSO, for optimal hyper-parameter selection in the transfer learning process with respect to HAR tasks with still images. It incorporates Gaussian fitness surface prediction and exponential search coefficients to overcome stagnation. It optimises the learning rate, batch size, and number of re-trained layers of a pre-trained convolutional neural network (CNN). To overcome bias of single optimised networks, an ensemble model with three optimised CNN streams is introduced. The first and second streams employ raw images and segmentation masks yielded by mask R-CNN as inputs, while the third stream fuses a pair of networks with raw image and saliency maps as inputs, respectively. The final prediction results are obtained by computing the average of class predictions from all three streams. By leveraging differences between learned representations within optimised streams, our ensemble model outperforms counterparts devised by PSO and other state-of-the-art methods for HAR. In addition, evaluated using diverse artificial landscape functions, EnvPSO performs better than other search methods with statistically significant difference in performance.

A generative adversarial network with joint multistream architecture and spectral compensation for pansharpening

Convolutional neural networks (CNNs) and variational models for pansharpening have obtained a compelling performance gain over the state of the art. Inspired by these models, we propose MSCGAN, a generative adversarial network (GAN) with joint multistream architecture and spectral compensation for pansharpening that uses a variational model to incorporate domain-specific knowledge and in particular, introduces a spectral compensation block. First, we extract the structural information of the panchromatic (PAN) image and input it into the generator together with the upsampled multispectral (MS) image. Then, we design a multistream pansharpening CNN architecture suite for domain-specific knowledge. Second, to boost the quality of the pansharpened images, we put the MS image in the generator and design a spectral compensation block. Then, we introduce the concept of the energy function of the variational model and add corresponding spectral constraints and spatial structure constraints to the objective function to achieve a compromise between spectral information fidelity and spatial information fidelity. Finally, the discriminator also introduces spatial structure information to help the generator generate the desired high-resolution multispectral image. Experiments on the images captured by the Quickbird and WorldView-2 satellites show that the our proposed MSCGAN can make use of PAN and LRMS images adequately to attain very encouraging results, obtaining large gains over the state of the art both visually and as measured by quality metrics.

Novel Architecture for Human Re-Identification with a Two-Stream Neural Network and Attention Mechanism

This paper proposes a novel architecture that utilises an attention mechanism in conjunction with multi-stream convolutional neural networks (CNN) to obtain high accuracy in human re-identification (Reid). The proposed architecture consists of four blocks. First, the pre-processing block prepares the input data and feeds it into a spatial-temporal two-stream CNN (STC) with two fusion points that extract the spatial-temporal features. Next, the spatial-temporal attentional LSTM block (STA) automatically fine-tunes the extracted features and assigns weight to the more critical frames in the video sequence by using an attention mechanism. Extensive experiments on four of the most popular datasets support our architecture. Finally, the results are compared with the state of the art, which shows the superiority of this approach.

Facial Beauty Prediction From Facial Parts Using Multi-Task and Multi-Stream Convolutional Neural Networks

Automatic analysis of facial beauty has become an emerging computer vision problem in recent years. Facial beauty prediction (FBP) aims at developing a human-like model that automatically makes facial attractiveness predictions. In this study, we present and evaluate a face attractiveness prediction approach using facial parts as well as a multi-task learning scheme. First, a deep convolutional neural network (CNN) pre-trained on massive face datasets is utilized for face attractiveness prediction, which is capable of automatic learning of high-level face representations. Next, we extend our deep model to other facial attribute recognition tasks. Hence, a multi-task learning scheme is leveraged by our deep model to learn optimal shared features for three correlated tasks (i.e. facial beauty assessment, gender recognition as well as ethnicity identification). To further enhance the attractiveness computation accuracy, specific regions of face images (i.e. left eye, nose and mouth) as well as the whole face are fed into multi-stream CNNs (i.e. three two-stream networks). Each two-stream network adopts a facial part as well as the full face as input. Extensive experiments are conducted on the SCUT-FBP5500 benchmark dataset, where our approach indicates significant improvement in accuracy over the other state-of-the-art methods.

Multi-view motion modelled deep attention networks (M2DA-Net) for video based sign language recognition

Currently, video-based Sign language recognition (SLR) has been extensively studied using deep learning models such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs). In addition, using multi view attention mechanism along with CNNs could be an appealing solution that can be considered in order to make the machine interpretation process immune to finger self-occlusions. The proposed multi stream CNN mixes spatial and motion modelled video sequences to create a low dimensional feature vector at multiple stages in the CNN pipeline. Hence, we solve the view invariance problem into a video classification problem using attention model CNNs. For superior network performance during training, the signs are learned through a motion attention network thus focusing on the parts that play a major role in generating a view based paired pooling using a trainable view pair pooling network (VPPN). The VPPN, pairs views to produce a maximally distributed discriminating features from all the views for an improved sign recognition. The results showed an increase in recognition accuracies on 2D video sign language datasets. Similar results were obtained on benchmark action datasets such as NTU RGB D, MuHAVi, WEIZMANN and NUMA as there is no multi view sign language dataset except ours.

Weighted voting of multi-stream convolutional neural networks for video-based action recognition using optical flow rhythms

Two of the most important premises of an ensemble are the diversity of its components and how to combine their votes. In this paper, we propose a multi-stream architecture based on the weighted voting of convolutional neural networks to deal with the problem of recognizing human actions in videos. A major challenge is how to include temporal aspects into this kind of approach. A key step in this direction is the selection of features that characterize the complexity of human actions in time. In this context, we propose a new stream, Optical Flow Rhythm, besides using other streams for diversity. To combine the streams, a voting system based on a new weighted average fusion method is introduced. In this scheme, the weights of classifiers are defined by an optimization process led by a metaheuristic. Experiments conducted on the UCF101 and HMDB51 datasets demonstrate that our method is comparable to state-of-the-art approaches.

Image Aesthetics Assessment Based on Multi-stream CNN Architecture and Saliency Features

ABSTRACT In this paper, we explore how higher-level perceptual information based on visual attention can be used for aesthetic assessment of images. We assume that visually dominant subjects in a photograph influence stronger aesthetic interest. In other words, visual attention may be a key to predicting photographic aesthetics. Our proposed aesthetic assessment method, which is based on multi-stream and multi-task convolutional neural networks (CNNs), extracts global features and saliency features from an input image. These provide higher-level visual information such as the quality of the photo subject and the subject–background relationship. Results from our experiments support the effectiveness of our approach.

Multi-view region-adaptive multi-temporal DMM and RGB action recognition

Human action recognition remains an important yet challenging task. This work proposes a novel action recognition system. It uses a novel multi-view region-adaptive multi-resolution-in-time depth motion map (MV-RAMDMM) formulation combined with appearance information. Multi-stream 3D convolutional neural networks (CNNs) are trained on the different views and time resolutions of the region-adaptive depth motion maps. Multiple views are synthesised to enhance the view invariance. The region-adaptive weights, based on localised motion, accentuate and differentiate parts of actions possessing faster motion. Dedicated 3D CNN streams for multi-time resolution appearance information are also included. These help to identify and differentiate between small object interactions. A pre-trained 3D-CNN is used here with fine-tuning for each stream along with multi-class support vector machines. Average score fusion is used on the output. The developed approach is capable of recognising both human action and human–object interaction. Three public-domain data-sets, namely MSR 3D Action, Northwestern UCLA multi-view actions and MSR 3D daily activity, are used to evaluate the proposed solution. The experimental results demonstrate the robustness of this approach compared with state-of-the-art algorithms.

L‐CNN: a lattice cross‐fusion strategy for multistream convolutional neural networks

This paper proposes a fusion strategy for multistream convolutional networks, the Lattice Cross Fusion. This approach crosses signals from convolution layers performing mathematical operation-based fusions right before pooling layers. Results on a purposely worsened CIFAR-10, a popular image classification data set, with a modified AlexNet-LCNN version show that this novel method outperforms by 46% the baseline single stream network, with faster convergence, stability, and robustness.

Region based multi-stream convolutional neural networks for collective activity recognition

Collective activity recognition, which analyses the behavior of groups of people in videos, is an important goal of video surveillance systems. In this paper, we focus on collective activity recognition problem and propose a new multi-stream convolutional neural network architecture that utilizes information extracted from multiple regions. The proposed method is the first work that uses a multi-stream network and multiple regions in this problem. Various strategies to fuse multiple spatial and temporal streams are explored. We evaluate the proposed method on two benchmark datasets, the Collective Activity Dataset and the Volleyball Dataset. Our experimental results show that the proposed method improves collective activity recognition performance when compared to the state-of-the-art approaches.

Parallel imaging in time-of-flight magnetic resonance angiography using deep multistream convolutional neural networks.

To develop and evaluate a method of parallel imaging time-of-flight (TOF) MRA using deep multistream convolutional neural networks (CNNs). A deep parallel imaging network ("DPI-net") was developed to reconstruct 3D multichannel MRA from undersampled data. It comprises 2 deep-learning networks: a network of multistream CNNs for extracting feature maps of multichannel images and a network of reconstruction CNNs for reconstructing images from the multistream network output feature maps. The images were evaluated using normalized root mean square error (NRMSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) values, and the visibility of blood vessels was assessed by measuring the vessel sharpness of middle and posterior cerebral arteries on axial maximum intensity projection (MIP) images. Vessel sharpness was compared using paired t tests, between DPI-net, 2 conventional parallel imaging methods (SAKE and ESPIRiT), and a deep-learning method (U-net). DPI-net showed superior performance in reconstructing vessel signals in both axial slices and MIP images for all reduction factors. This was supported by the quantitative metrics, with DPI-net showing the lowest NRMSE, the highest PSNR and SSIM (except R = 3.8 on sagittal MIP images, and R = 5.7 on axial slices and sagittal MIP images), and significantly higher vessel sharpness values than the other methods. DPI-net was effective in reconstructing 3D TOF MRA from highly undersampled multichannel MR data, achieving superior performance, both quantitatively and qualitatively, over conventional parallel imaging and other deep-learning methods.

Cervical Cell Features Based Multi-Stream Convolutional Neural Networks Classification Method

Cervical Cell Features Based Multi-Stream Convolutional Neural Networks Classification Method

Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

Despite the fact that automatic target recognition (ATR) in Synthetic aperture radar (SAR) images has been extensively researched due to its practical use in both military and civil applications, it remains an unsolved problem. The major challenges of ATR in SAR stem from severe data scarcity and great variation of SAR images. Recent work started to adopt convolutional neural networks (CNNs), which, however, remain unable to handle the aforementioned challenges due to their high dependency on large quantities of data. In this paper, we propose a novel deep convolutional learning architecture, called Multi-Stream CNN (MS-CNN), for ATR in SAR by leveraging SAR images from multiple views. Specifically, we deploy a multi-input architecture that fuses information from multiple views of the same target in different aspects; therefore, the elaborated multi-view design of MS-CNN enables it to make full use of limited SAR image data to improve recognition performance. We design a Fourier feature fusion framework derived from kernel approximation based on random Fourier features which allows us to unravel the highly nonlinear relationship between images and classes. More importantly, MS-CNN is qualified with the desired characteristic of easy and quick manoeuvrability in real SAR ATR scenarios, because it only needs to acquire real-time GPS information from airborne SAR to calculate aspect differences used for constructing testing samples. The effectiveness and generalization ability of MS-CNN have been demonstrated by extensive experiments under both the Standard Operating Condition (SOC) and Extended Operating Condition (EOC) on the MSTAR dataset. Experimental results have shown that our proposed MS-CNN can achieve high recognition rates and outperform other state-of-the-art ATR methods.

Action Recognition by an Attention-Aware Temporal Weighted Convolutional Neural Network.

Research in human action recognition has accelerated significantly since the introduction of powerful machine learning tools such as Convolutional Neural Networks (CNNs). However, effective and efficient methods for incorporation of temporal information into CNNs are still being actively explored in the recent literature. Motivated by the popular recurrent attention models in the research area of natural language processing, we propose the Attention-aware Temporal Weighted CNN (ATW CNN) for action recognition in videos, which embeds a visual attention model into a temporal weighted multi-stream CNN. This attention model is simply implemented as temporal weighting yet it effectively boosts the recognition performance of video representations. Besides, each stream in the proposed ATW CNN framework is capable of end-to-end training, with both network parameters and temporal weights optimized by stochastic gradient descent (SGD) with back-propagation. Our experimental results on the UCF-101 and HMDB-51 datasets showed that the proposed attention mechanism contributes substantially to the performance gains with the more discriminative snippets by focusing on more relevant video segments.

Deep Learning and Multi-Sensor Fusion for Glioma Classification Using Multistream 2D Convolutional Networks.

This paper addresses issues of brain tumor, glioma, grading from multi-sensor images. Different types of scanners (or sensors) like enhanced T1-MRI, T2-MRI and FLAIR, show different contrast and are sensitive to different brain tissues and fluid regions. Most existing works use 3D brain images from single sensor. In this paper, we propose a novel multistream deep Convolutional Neural Network (CNN) architecture that extracts and fuses the features from multiple sensors for glioma tumor grading/subcategory grading. The main contributions of the paper are: (a) propose a novel multistream deep CNN architecture for glioma grading; (b) apply sensor fusion from T1-MRI, T2-MRI and/or FLAIR for enhancing performance through feature aggregation; (c) mitigate overfitting by using 2D brain image slices in combination with 2D image augmentation. Two datasets were used for our experiments, one for classifying low/high grade gliomas, another for classifying glioma with/without 1p19q codeletion. Experiments using the proposed scheme have shown good results (with test accuracy of 90.87% for former case, and 89.39 % for the latter case). Comparisons with several existing methods have provided further support to the proposed scheme. keywords: brain tumor classification, glioma, 1p19q codeletion, glioma grading, deep learning, multi-stream convolutional neural networks, sensor fusion, T1-MR image, T2-MR image, FLAIR.