In this paper we study two-layered heterogeneous sensor networks where two types of nodes are deployed in the network: basic and cluster head nodes. Basic simple inexpensive, while much powerful richer energy. A node organizes around it into a cluster. does data collections sends packets when polled by head. By introducing hierarchy, such network has better scalability than homogeneous networks. It also smaller overall cost since networking functionalities shifted from sensors to longer life...

Short-term hourly reliable prediction of significant wave height is an important research topic in coastal engineering. Many researchers have carried out in-depth studies many ocean regions. Generally, most this work implemented through numerical models. However, as for models, with the increase duration, accumulation randomness leads to poor effect. In paper, four buoy stations Taiwan Strait are taken objects. We propose a algorithm, which combines weather model WRF and deep-learning model,...

Semi-supervised learning is a pattern that can utilize labeled data and unlabeled to train deep neural networks. In semi-supervised methods, self-training-based methods do not depend on augmentation strategy have better generalization ability. However, their performance limited by the accuracy of predicted pseudo-labels. this paper, we propose reduce noise in pseudo-labels from two aspects: predictions confidence predictions. For first aspect, similarity graph structure (SGSL) model...

Mental fatigue is a state that may occur due to excessive work or long-term stress. Electroencephalography (EEG) considered reliable standard for mental detection. The existing EEG detection methods mainly use traditional machine learning models classify after manual feature extraction. However, extraction difficult and complicated. quality of largely determines the model. In this article, we collected signals from 30 medical staff. wavelet threshold denoising method was then applied...

Aiming at the problem that edge points are difficult to be accurately divided in DBSCAN algorithm, a density clustering algorithm based on silhouette coefficient constraints (CDBSCAN) is proposed. The CDBSCAN improves accuracy with as criterion. Firstly, data preliminary classified by and number of formed clusters calculated. Then both noise fewer listed potential data. Subsequently, each point set again according coefficient. Finally, experiments conducted synthetic public datasets, result...

All-optical communication, in particular, wavelength-division-multiplexing (WDM) technique, has been proposed as a promising candidate to meet the ever-increasing demands on bandwidth from emerging bandwidth-intensive computing/networking applications. However, with current technology, cost of optical especially buffering and wavelength conversion, remains major concern for such In this paper, we study WDM interconnects that utilize low recirculating limited range conversion. We first...

Mental fatigue is a state that may occur due to excessive work or long-term stress. Electroencephalography (EEG) considered reliable standard for mental detection. The existing EEG detection methods mainly use traditional machine learning models classify after manual feature extraction. However, extraction difficult and complicated. quality of largely determines the model. In this article, we collected signals from 30 medical staff. wavelet threshold denoising method was then applied...

Multicast involves transmitting information from a single source to multiple destinations, and is an important operation in high-performance networks. A k-fold multicast network was recently proposed as cost-effective solution providing better quality-of-service functions supporting real-world applications. To give quantitative basis for designers determine the suitable value of system parameter k under different traffic loads, this paper, we propose analytical model performance networks...

Wave forecasting approaches based on deep learning techniques have recently made a great progress. In this study, we developed model Gated Recurrent Unit (GRU) and sequence-to-sequence neural networks (GRUS), to improve the accuracy of significant wave heights for Taiwan Strait, where ocean waves winds own their unique characteristics. The performances our proposed GRUS other models WaveNet Long Short-Term Memory (LSTM) were compared by means wind observations at three buoys in study area....

In spite of great success in many visual recognition tasks achieved by recent deep models, they performed poorly at low resolution conditions. low-resolution image is still a challenging problem. The massive loss detail information the ultimate cause this However, individual instance details has little effect on relationship between instances.Therefore, we propose multi-level relational distillation method to solve identification Based teacher student framework knowledge distillation, It...

The trajectory of oil spill or missing person on the sea surface marks movement seawater, so forecasting drift floating object in open ocean is an important application. Many nations support services for search-and-rescue and ship safety, all which may benefit from forecasts. We implement a marine prediction system. Take advantage ROMS (Regional Ocean Model System) model forced by meteorology wind field, our can provide possible trajectories drifting object. Therefore, short time after...

The presence of tortuosity in the retinal vessels is crucial diagnosis ocular fundus disorders. There are numerous methods for computing arteries available today, all which have yielded impressive results. However, they usually divide into smaller vascular structures to calculate local tortuosities, then weighted summed get global entire vessel. approach division on a two-dimensional image weakens vessel information and makes it unable accurately portray vessel's tortuosity. Hence, we...

Majority voting outlier detection is a traditional method that has been widely used in many fields. It uses the strategy of majority vote to make prediction, which makes it perform poorly acc index sometimes. In this paper, called second anomaly (SAD) proposed, detect connection scores between each other and decide advantage strength sample when defining outlierness, expressed as <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$a$</tex> factor,...