This paper has proposed an architecture of optimised SIFT (scale invariant feature transform) detection for FPGA implementation image matcher. In order based matcher to be implemented on efficiently, in terms speed and hardware resource usage, the original algorithm been significantly following aspects: 1) upsampling replaced with downsampling save interpolation operation. 2) Only four scales two octaves are needed our moderate degradation matching performance. 3) The total dimension...

Deep convolutional neural networks have achieved great success on image classification. A series of feature extractors learned from CNN been used in many computer vision tasks. Global pooling layer plays a very important role deep networks. It is found that the input feature-maps global become sparse, as increasing use Batch Normalization and ReLU combination, which makes original low efficiency. In this paper, we proposed novel end-to-end trainable operator AlphaMEX Pool for network....

Step length estimation is an important issue in areas such as gait analysis, sport training, or pedestrian localization. In this article, we estimate the step of walking using a waist-worn wearable computer named eButton. Motion sensors within device are used to record body movement from trunk instead extremities. Two signal-processing techniques applied our algorithm design. The direction cosine matrix transforms vertical acceleration coordinates topocentric coordinates. empirical mode...

Convolutional neural networks (CNNs) are becoming more and popular today. CNNs now have become a feature extractor applying to image processing, big data fog computing, etc. usually consist of several basic units like convolutional unit, pooling activation so on. In CNNs, conventional methods refer 2×2 max‐pooling average‐pooling, which applied after the or ReLU layers. this paper, we propose Multiactivation Pooling (MAP) Method make accurate on classification tasks without increasing depth...

Due to the flexibility and ease of deployment Field Programmable Gate Arrays (FPGA), more studies have been conducted on developing optimizing target detection algorithms based Convolutional Neural Networks (CNN) models using FPGAs. Still, these focus improving performance core algorithm hardware structure, with few focusing unified architecture design corresponding optimization techniques for model, resulting in inefficient overall model performance. The essential reason is that do not...

We present a straightforward brightness preserving image enhancement technique. The proposed method is based on an original gradient and intensity histogram (GIH) which contains both information of the image. This character enables GIH to avoid high peaks in traditional and, thus alleviate overenhancement our method, i.e., equalization (GIHE). GIHE can also enhance strength image, good for improving subjective quality since human vision system more sensitive than absolute Considering that...

The emergence of new wearable technologies, such as action cameras and smart glasses, has driven the use first-person perspective in computer applications. This field is now attracting attention investment researchers aiming to develop methods process vision (FPV) video. current approaches present particular combinations different image features quantitative accomplish specific objectives, object detection, activity recognition, user–machine interaction, etc. FPV-based navigation necessary...

Convolutional neural networks (CNN) are mainly used for image recognition tasks. However, some huge models infeasible mobile devices because of limited computing and memory resources. In this paper, feature maps DenseNet CondenseNet visualized. It could be observed that there channels in locked state have similar distribution property, which compressed further. Thus, work, a novel architecture — RSNet is introduced to improve the efficiency CNNs. This paper proposes Relative-Squeezing (RS)...

Federated Learning (FL), a promising deep learning paradigm extensively deployed in Vehicular Edge Computing Networks (VECN), allows distributed approach to train datasets of nodes locally, e.g., for mobile vehicles, and exchanges model parameters obtain an accurate without raw data transmission. However, the existence malicious vehicular as well inherent heterogeneity vehicles hinders attainment models. Moreover, local training parameter transmission during FL exert notable energy burden on...

In the object detection task, deep learning-based methods always need a large amount of annotated training data. However, annotating number images is labor-intensive. order to reduce dependency expensive annotations, we propose novel end-to-end feature reconstruction and metric based network for few-shot (FM-FSOD). FM-FSOD integrates learning meta-learning tackle task. class-agnostic model that can accurately recognize categories without fine-tuning on categories. Specifically, quickly learn...

The continuous wave interferences (CWIs) and the narrow-band (NBIs) have significantly impacted acquisition, tracking positioning accuracy of Beidou Navigation Satellite System (BDS) receivers. As an interference suppression technology with a simple structure low hardware cost, adaptive infinite-duration impulse response (IIR) notch filter has been widely employed in receivers to mitigate CWIs NBIs. However, nonlinear phase characteristics introduced by IIR filters into navigation receivers,...

In the object detection task, CNN (Convolutional neural networks) models always need a large amount of annotated examples in training process. To reduce dependency expensive annotations, few-shot has become an increasing research focus. this paper, we present effective framework (MM-FSOD) that integrates metric learning and meta-learning to tackle task. Our model is class-agnostic can accurately recognize new categories, which are not appearing samples. Specifically, fast learn features...

The primary bottleneck of many model-based diagnosis approaches is computing minimal hitting sets (MHSs). Most existing algorithms for MHSs are deterministic. Such sound and complete but with the complexity exponential, since problem NP-hard. To reduce complexity, especially large systems, we present a novel approach multiple fault diagnosis, based on an immune genetic algorithm (IGA). This maps set onto integer programming problem, followed by obtaining lower bound size solution. use in...

The Costas loop is the most common carrier synchronization structure in receiver using BPSK modulation. Reliable detection of lock condition important to receiver. In this paper, an improved algorithm for proposed. First, mathematical model and hardware are proposed, then theoretical derivation given. Through analysis performance, criterion choosing threshold presented. Simulation results show that signal power normalization can eliminate impact input changes. work reliably when Es/N0> 1dB,...

Satellite signal anomalies result from a failure of the generating hardware on one GNSS space vehicles. These may cause severe distortions autocorrelation peak inside receivers, which could threaten position accuracy and integrity GNSS, especially for safety critical applications. Digital distortion caused by failures baseband generation unit is an important kind deformations. Multi-correlator technique, as prevalent Signal Quality Monitoring (SQM) method, has shown good performance...

Abstract This paper investigates long-term visual object tracking which is a complex problem in computer vision community and big data analysis, due to the variation of target surrounding environment. A novel algorithm based on local correlation filtering global keypoint matching proposed solve problems occurred during such as occlusion, target-losing, etc. The consists two major components: (1) module, (2) losing re-detection module. module optimizes conventional algorithm. Firstly, Color...

Convolutional neural networks(CNN) are showing powerful performance on image recognition tasks. However, when CNN is applied to mobile devices, with limited computing and memory resource, it requires more compact design maintain a relatively high performance. In this paper, we propose Relative Squeezing Net(RSNet) that provides technical insight into structure for designing model. an endeavor improve CondenseNet, introduce Relative-Squeezing bottleneck where output weighted percentage of...

Using longer pseudo-random code in spread spectrum communication system can provide many benefits, such as improving the security of communication, increasing ambiguous distance pseudo ranging. But it has comparatively difficulty acquisition. In this paper, disadvantages traditional FFT acquisition algorithm are analyzed firstly. The correlation peak affects at low signal to noise ratio under influence initial phase code. To problem, an improved fast is proposed, which uses two channels...

Inter-satellite links are an important component of the new generation satellite navigation systems, characterized by low signal-to-noise ratio (SNR), complex electromagnetic interference and short time slot each satellite, which brings difficulties to acquisition stage. The inter-satellite link in both Global Positioning System (GPS) BeiDou Navigation Satellite (BDS) adopt long code spread spectrum system. However, is a difficult time-consuming task due period. Traditional folding methods...

This study investigates the use of a chest-worn wearable computer, eButton, to assess physical performance older adults. The Short Physical Performance Battery (SPPB), standard cliniucal test, is first conducted on human subjects. Then, triaxial accelerometer and gyroscope within eButton are utilized record acceleration angular velocity body motion same subjects for one week. sensor data corresponding walking episodes segmented features in time frequency domains extracted. Comparison between...

The large Doppler offset that exists in high dynamic environments poses a serious impediment to the acquisition of direct sequence spread spectrum (DSSS) signals. To ensure acceptable detection probabilities, frequency space has be finely divided, which leads complicated structures and excessively long time at low SNR. A local folding (LFF) method designed for combined application with established techniques dedicated PN-code synchronization is proposed this paper. Through modulating block...