Automatic modulation classification (AMC) is the key technique in both military and civilian wireless communication. However, performance unsatisfactory, even several deep learning-based methods are involved. Targeting its low accuracy at SNR, high computational cost label overdependence, we propose a novel AMC framework, where autoencoder (AE) serves as backbone Convolution-AE LSTM-AE combined parallel way temporal spatial feature extractors. The comparisons with serval algorithms on...

Background: Thioredoxin 1 (Trx-1) is a small redox protein predominantly localized in the cytoplasm. Its expression increased several cancers, including colorectal cancer (CRC). However, function of Trx-1 translocation to nucleus not clear. In this study, we investigated role nuclear development CRC. Methods: Expression and STAT3 was analyzed by Western blot immunofluorescence. Endogenous interaction Trx-1, STAT3, karyopherin α1 CRC cells co-immunoprecipitation. pSTAT3 staining human tissues...

This article presents a structure constraint matrix factorization framework for different behavior segmentation of the human sequential data. is based on structural information continuity and high similarity between neighboring frames. Due to dimensionality data, high-precision hard achieve from perspective application academia. By making hypothesis, first, effective regular terms are constructed. Subsequently, clustering constrained non-negative established. Finally, result can be obtained...

The Segment Anything Model (SAM) represents a state-of-the-art research advancement in natural image segmentation, achieving impressive results with input prompts such as points and bounding boxes. However, our evaluation recent indicate that directly applying the pretrained SAM to medical segmentation does not yield satisfactory performance. This limitation primarily arises from significant domain gap between images images. To bridge this gap, we introduce SAM-Med2D, most comprehensive...

The carbon footprint associated with large language models (LLMs) is a significant concern, encompassing emissions from their training, inference, experimentation, and storage processes, including operational embodied emissions. An essential aspect accurately estimating the impact of emerging LLMs even before which heavily relies on GPU usage. Existing studies have reported LLM but only one tool, mlco2, can predict new neural networks prior to physical training. However, mlco2 has several...

Segment Anything Model (SAM) has achieved impressive results for natural image segmentation with input prompts such as points and bounding boxes. Its success largely owes to massive labeled training data. However, directly applying SAM medical cannot perform well because lacks knowledge -- it does not use images training. To incorporate into SAM, we introduce SA-Med2D-20M, a large-scale dataset of 2D built upon numerous public private datasets. It consists 4.6 million 19.7 corresponding...

Face recognition has attracted great interest especially in computer vision, but poses and illumination of the face lay difficulty accuracy. Studies on human field cognitive neuroscience found some face-specific event-related potential (ERP) components, such as N170, VPP. And researchers used these ERP components to achieve detection based Brain Computer Interface (BCI) system. In this paper, we introduce a BCI system using rapid serial visual presentation (RSVP) paradigm evaluate...

Current multibeam satellite systems perform centralized power and carrier allocation assuming a multi-frequency time division multiple access (MF-TDMA) scheme with fixed regular frequency reuse. This approach is not optimal when individual SINR constraints are to be met. In this paper, we formulate the problem of joint allowing non-regular setting will increase interference levels but in turn allow more efficient use increasing spectral efficiency. We focus on distributed formulation compare...

This paper proposes a relay channel model which takes into account the influence of location and antenna configuration. Simulation results show that when is in line-of-sight (LOS) propagation environment, capacity will degrade. To improve performance channel, water-filling (WF) algorithm used to allocate power, simulation shows WF-based outperforms uniform power allocation channel. The radiation pattern also an importance factor affects this case. widely Rayleigh overestimates compared with...

In this paper, we study the channel estimation problem for time division multiplex amplify-and-forward (AF) relay channels. Both source and destination also in-between node comprise multiple antennas, there are spatial fading correlations in each hop. We first derive a linear minimum mean square error (LMMSE) estimator from simple scenario where no among antennas to more general spatially correlated. And then design optimal training sequences by minimizing (MSE). Also impacts of...

Recent face recognition techniques have achieved remarkable successes in fast retrieval on huge image datasets. But the performance is still limited when large illumination, pose, and facial expression variations are presented. In contrast, human brain has powerful cognitive capability to recognize faces demonstrates robustness across viewpoints, lighting conditions, even presence of partial occlusion. This paper proposes a closed-loop system that combines state-of-the-art method with...

In modulation classification domain, handcrafted feature based method can fit well from a few labeled samples, while deep learning require large amount of samples to achieve the superior performance. order improve accuracy under constraint limited this paper proposes few-shot on dimension reduction and pseudo-label training (FDRPLT), which combines with method. First, an optimal low-dimensional subset is created by combination features autoencoder-extracted post-processed selection...

Perception and navigation control is core key technology for the legged robot to adapt complex terrain achieve autonomous walking. It distinguish from manned equipment become a ground unmanned system. In this paper, SLAM globally localization system based on lidar point cloud established, framework combination of topological measurements proposed realize map reconstruction wild environment. Path planning Dijkstra algorithm implement task quadruped robot, obstacle avoidance strategy local...

The automatic machine face recognition has achieved great performance but is still far from satisfactory in the uncontrolled environments. Human brain a powerful ability to recognize faces across various conditions, which makes it possible introduce brain-computer interface (BCI) technology recognition. However, of single-participant based BCI suffers low signal-to-noise ratio electroencephalography (EEG) signals, especially detecting event-related potential (ERP) single trial. Here, we...

The capacity of M-ary PPM ultra-wideband communications over multipath channels is analyzed in this paper. And the closed form output signal-to-noise ratio (SNR) receiver derived. In addition, link budget taken into account incorporate to specific constraints, which are mainly power spectrum density limitation under FCC part 15 rules. Finally, influence muitipath number on and compared user number. It shown that decreased faster with increasement than

(Abstract) In this paper, the transport efficiency of Multi Protocol Encapsulation (MPE), Unidirectional Lightweight (ULE) and Generic Stream (GSE) for typical IP packet sizes is compared. Moreover, aggregated when applying packet-level forward error correction (PL-FEC) with MPE, ULE GSE also analyzed. MPE-FEC mechanism used by DVB-H whereas GSE-FEC our proposed modification to be in DVB-S2. A layered calculation model presented order simplify computation. The performance analyzed adopted...

Based on end effects of empirical mode decomposition (EMD) Hilbert-Huang Transform (HHT) in detecting ultra-wideband (UWB) signal, the non-equidistance grey model (NGM) is analyzed to mitigate EMD by predicting uncertainly data. Since some extreme points can hardly be detected particular situation, modified NGM(1,1)model using Fourier series (TFNGM(1,1)) at time domain proposed. Furthermore, new criterion terminate iteration process also proposed and this paper, which effectively improve...

A novel approach named adaptive modified receiving template waveform (AMRTW) is proposed in this paper. This new uses frequency estimator to give the estimating values of interference and band. Based on these values, signal decomposed rationally by signal-analyzing unit. And UWB receiver modifies with removing some decomposing components that are close interfering spectrum from signal. The simulation results bit error rate performance show method efficient mitigation for pulse radio.

The event-related potential (ERP) is the brain response measured in electroencephalography (EEG), which reflects process of human cognitive activity. ERP has been introduced into computer interfaces (BCIs) to communicate with subject's intention. Due low signal-to-noise ratio EEG, most studies are based on grand-averaging over many trials. Recently single-trial detection attracts more attention, enables real time processing tasks as rapid face identification. All targets needed be retrieved...

Computer vision (CV) has demonstrated its capabilities of face retrieval in rapidly exploring large image collections, but performance degrades uncontrolled condition. Human brain powerful cognitive ability to recognize faces across various conditions, which could complement computer retrieval. In this paper, we propose a closed-loop system by iterative coupling between EEG-based target labeling and CV based label propagation. The starts with sample the sent module. module ranks all images...

In this paper, a blind channel estimator, which is based on linear non-redundant precoding and second-order statistics, proposed for space time block (STBC) coded MIMO-OFDM systems. According to the complex orthogonal properties of STBCs, estimate with scalar ambiguity can be obtained through series simple transformation. Simulation results show that scheme track time-varying state well thus improve system performance when number statistical data properly chosen.

Stroke is one of the leading causes mortality and disability worldwide, with limited therapeutic approaches. As an endogenous strategy for neuroprotection, postconditioning treatments have proven to be promising therapies against cerebral ischemia. However, complicated procedures potential safety issues limit their clinical application. To overcome these disadvantages, we developed acidic (APC) as a therapy experimental focal APC refers mild acidosis treatment by inhaling CO2 during...

Software-defined networking (SDN) has recently been all the rage. It already successfully applied in many practical fields including campus network, wide area network (WAN) and data centre network. As a next step, SDN is expected to be adopted cellular networks deal with some most urgent problems faced by carriers, such as ever soaring capacity demands differentiated quality of service (QoS)/quality experience (QoE) requirements introduced emerging services large number exciting...