Artificial intelligence (AI) has incorporated various automatic systems and frameworks to diagnose the severity of depression using hand-crafted features. However, process of feature selection needs ...domain knowledge and is still time-consuming and subjective. Deep learning technology has been successfully adopted for depression recognition. Most previous works pre-train the deep models on large databases followed by fine-tuning with depression databases (i.e., AVEC2013, AVEC2014). In the present paper we propose an integrated framework – Deep Local Global Attention Convolutional Neural Network (DLGA-CNN) for depression recognition, which adopts CNN with attention mechanism as well as weighted spatial pyramid pooling (WSPP) to learn a deep and global representation. Two branches are introduced: Local Attention based CNN (LA-CNN) focuses on the local patches, while Global Attention based CNN (GA-CNN) learns the global patterns from the entire facial region. To capture the complementary information between the two branches, Local–Global Attention-based CNN (LGA-CNN) is proposed. After feature aggregation, WSPP is used to learn the depression patterns. Comprehensive experiments on AVEC2013 and AVEC2014 depression databases have demonstrated that the proposed method is capable of mining the underlying depression patterns of facial videos and outperforms the most of the state-of-the-art video-based depression recognition approaches.
Manual segmentation of the brain tumors for cancer diagnosis from MRI images is a difficult, tedious, and time-consuming task. The accuracy and the robustness of brain tumor segmentation, therefore, ...are crucial for the diagnosis, treatment planning, and treatment outcome evaluation. Mostly, the automatic brain tumor segmentation methods use hand designed features. Similarly, traditional methods of deep learning such as convolutional neural networks require a large amount of annotated data to learn from, which is often difficult to obtain in the medical domain. Here, we describe a new model two-pathway-group CNN architecture for brain tumor segmentation, which exploits local features and global contextual features simultaneously. This model enforces equivariance in the two-pathway CNN model to reduce instabilities and overfitting parameter sharing. Finally, we embed the cascade architecture into two-pathway-group CNN in which the output of a basic CNN is treated as an additional source and concatenated at the last layer. Validation of the model on BRATS2013 and BRATS2015 data sets revealed that embedding of a group CNN into a two pathway architecture improved the overall performance over the currently published state-of-the-art while computational complexity remains attractive.
Hyperspectral image (HSI) classification is widely used for the analysis of remotely sensed images. Hyperspectral imagery includes varying bands of images. Convolutional neural network (CNN) is one ...of the most frequently used deep learning-based methods for visual data processing. The use of CNN for HSI classification is also visible in recent works. These approaches are mostly based on 2-D CNN. On the other hand, the HSI classification performance is highly dependent on both spatial and spectral information. Very few methods have used the 3-D-CNN because of increased computational complexity. This letter proposes a hybrid spectral CNN (HybridSN) for HSI classification. In general, the HybridSN is a spectral-spatial 3-D-CNN followed by spatial 2-D-CNN. The 3-D-CNN facilitates the joint spatial-spectral feature representation from a stack of spectral bands. The 2-D-CNN on top of the 3-D-CNN further learns more abstract-level spatial representation. Moreover, the use of hybrid CNNs reduces the complexity of the model compared to the use of 3-D-CNN alone. To test the performance of this hybrid approach, very rigorous HSI classification experiments are performed over Indian Pines, University of Pavia, and Salinas Scene remote sensing data sets. The results are compared with the state-of-the-art hand-crafted as well as end-to-end deep learning-based methods. A very satisfactory performance is obtained using the proposed HybridSN for HSI classification. The source code can be found at https://github.com/gokriznastic/HybridSN.
Plant pathologists desire soft computing technology for accurate and reliable diagnosis of plant diseases. In this study, we propose an efficient soybean disease identification method based on a ...transfer learning approach by using a pre-trained convolutional neural network (CNN’s) such as AlexNet, GoogleNet, VGG16, ResNet101, and DensNet201. The proposed convolutional neural networks were trained using 1200 plant village image dataset of diseased and healthy soybean leaves, to identify three soybean diseases out of healthy leaves. Pre-trained CNN used to enable a fast and easy system implementation in practice. We used the five-fold cross-validation strategy to analyze the performance of networks. In this study, we used a pre-trained convolutional neural network as feature extractors and classifiers. The experimental results based on the proposed approach using pre-trained AlexNet, GoogleNet, VGG16, ResNet101, and DensNet201 networks achieve an accuracy of 95%, 96.4 %, 96.4 %, 92.1%, 93.6% respectively. The experimental results for the identification of soybean diseases indicated that the proposed networks model achieves the highest accuracy
Physiotherapy exercises like extension, flexion, and rotation are an absolute necessity for patients of post stroke rehabilitation (PSR). A physiotherapist uses many techniques to restore movements ...needs in daily life including nerve re-education, task training, muscle strengthening and uses various assistive techniques. But, a physiotherapist guiding the physiotherapy exercises to a patient is a time-consuming, tedious and costly affair. In the paper, a novel automated system is designed for detecting and recognizing upper limb exercises using an RGB-Depth camera that could guide the patients to perform real-time physiotherapy exercises without human intervention. Hybrid deep learning (HDL) approaches are exploited for the highly accurate and robust system for recognizing physiotherapy exercises of the upper limb for PSR. As a baseline, a deep convolutional neural network (CNN) is designed that automatically extracts features from the pre-processed data and classifies the performed physiotherapy exercise. As the exercise is being performed, to extract and utilize temporal dependencies, architectures of recurrent neural network (RNN) are used. In the CNN-LSTM model, CNN derives useful features that are provided to LSTM thus increasing the accuracy of recognized exercises. To train faster, another hybrid deep learning model, CNN-GRU is implemented where a novel focal loss criterion is used to overcome the drawbacks of standard cross-entropy loss. Experimental evaluation is done using RGB-D data obtained from Microsoft Kinect v2 sensors. Dataset comprising of 10 different physiotherapy exercises were created. Experimental results have shown significant activity recognition accuracy with 98% and 99% for CNN and CNN-LSTM model respectively. CNN-GRU model is the best suitable architecture with 100% accuracy.
Evolutionary neural architecture search (ENAS) can automatically design the architectures of deep neural networks (DNNs) using evolutionary computation algorithms. However, most ENAS algorithms ...require an intensive computational resource, which is not necessarily available to the users interested. Performance predictors are a type of regression models which can assist to accomplish the search, while without exerting much computational resource. Despite various performance predictors have been designed, they employ the same training protocol to build the regression models: 1) sampling a set of DNNs with performance as the training dataset; 2) training the model with the mean square error criterion; and 3) predicting the performance of DNNs newly generated during the ENAS. In this article, we point out that the three steps constituting the training protocol are not well thought-out through intuitive and illustrative examples. Furthermore, we propose a new training protocol to address these issues, consisting of designing a pairwise ranking indicator to construct the training target, proposing to use the logistic regression to fit the training samples, and developing a differential method to build the training instances. To verify the effectiveness of the proposed training protocol, four widely used regression models in the field of machine learning have been chosen to perform the comparisons on two benchmark datasets. The experimental results of all the comparisons demonstrate that the proposed training protocol can significantly improve the performance prediction accuracy against the traditional training protocols.
This paper presents a new mechanism which is more effective for wearable devices to classify patient-specific electrocardiogram (ECG) heartbeats. In our method, a Generic Convolutional Neural Network ...(GCNN) is trained first using a large number of heartbeats without distinguishing patients. Based on the GCNN, fine-tuning technique is applied to modify the GCNN to a Tuned Dedicated CNN (TDCNN) for the corresponding individual. Notably, only the GCNN instead of common training data is required to be stored into wearable devices. Moreover, only fine-tuning with several seconds rather than dozens of minutes is needed before the TDCNN is used to monitor the long-term ECG signals in clinical. To accelerate the ECG classification, only the original ECG heartbeat is input to the CNN without other extended information from the neighbor heartbeats or FFT representation. A deeper CNN architecture with small-scale convolutional kernels is adopted to improve the speed and accuracy for classification. With deeper CNN, hierarchical features can be extracted to help improve the accuracy of ECG classification. The state-of-the-art performance on efficiency and accuracy for ECG classification over MIT-BIH dataset is achieved by the proposed method. The effectiveness and superiority for detecting ventricular ectopic beats (VEB) and supraventricular ectopic beats (SVEB) events are demonstrated. The proposed mechanism of fine-tuning the GCNN to TDCNN improves the efficiency for training patient-specific CNN classifier. Because of the computational efficiency of fine-tuning, ECG diagnosis and heart monitoring can be easily implemented with popular wearable devices in practice.
Deep convolutional networks are vulnerable to image translation or shift, partly due to common down-sampling layers, e.g., max-pooling and strided convolution. These operations violate the Nyquist ...sampling rate and cause aliasing. The textbook solution is low-pass filtering (blurring) before down-sampling, which can benefit deep networks as well. Even so, non-linearity units, such as ReLU, often re-introduce the problem, suggesting that blurring alone may not suffice. In this work, first, we analyse deep features with Fourier transform and show that Depth Adaptive Blurring is more effective, as opposed to monotonic blurring. To this end, we propose a novel Depth Adaptive Blur-pool (DAB-pool) module to replace existing down-sampling methods. Second, we introduce a novel activation function – with a built-in low pass filter, as an additional measure, to keep the problem from reappearing. From experiments, we observe generalisation on other forms of transformations and corruptions as well, e.g., rotation, scale, and noise. We evaluate our method under three challenging settings: (1) a variety of image translations; (2) adversarial attacks – both ℓp bounded and unbounded; and (3) data corruptions and perturbations. In each setting, our method achieves state-of-the-art results and improves clean accuracy on various benchmark datasets.
•A Dual-Stream Network (DSN) is developed for Photovoltaic Power Forecasting (PPF).•DSN extracts spatiotemporal features from actual historical data parallelly.•A self-attention mechanism is used to ...select optimal features.•Extensive experiments are performed to the choose best optimal model for PPF.•The DSN achieved better performance compared to baselines.
The operations of renewable power generation systems highly depend on precise Photovoltaic (PV) power forecasting, providing significant economic, and environmental advantages for energy efficient buildings and urban energy systems. However, precise PV power forecasting, particularly, solar power is more challenging due to solar energy intermittence, instability, and randomness. These challenges hinder the integration of PV into smart grids, where accurate power forecasting is a promising solution in this direction, providing effective planning and management services. Therefore, in this work, we introduce a dual-stream network for accurate PV forecasting. The proposed network parallelly learns spatial patterns using convolutional network and temporal representations via sequential learning algorithm. These features are then integrated together to form a single, yet representative feature vector used as an input to self-attention mechanism to further select the optimal features for PV power forecasting. To the best of our knowledge, the proposed dual stream network with advanced features selection mechanism is a pioneering approach for time series analysis, narrowed towards PV power forecasting. We derive our network after a series of experimentations involving solo and hybrid models, resulting in higher forecasting accuracy against state-of-the-art models.