Search

Soccer video event detection is of great significance to video retrieval.However,there are fewer events in soccer videos,and these events mainly occur in the far-view shot,which makes it difficult to capture key players and key actions,making soccer event detection more difficult.In recent years,methods based on deep learning have made some progress in soccer video event detection,but the learning ability of the high-level semantic of the event is still insufficient and the detection results need to be further improved.Therefore,how to improve the accuracy of soccer video event detection is an urgent problem to be solved.Taking the shooting event of free kick(free-kick shot event) as the research object,an event detection model combining soccer rules and deep learning is proposed.To have a deeper understanding of the inherent characteristics of the free-kick shot event,the event rules are manually summarized and verified on the public soccer dataset,and the corresponding application scenarios are also proposed.For the problem of too few events in soccer videos,rule-based initial localization algorithm is proposed to preprocess the videos.Through the combination and application of multiple rules,the location where the free-kick shot event may occur is initially located from the original video,which is used as the input of the deep learning model for further prediction.The proposed mo-del is compared with other models on the public soccer dataset.Experimental results show that the proposed model achieves the best results,with the accuracy rate of 78% and the recall rate of 81.25%.Compared with other models,the improvement in accuracy is particularly prominent.It can be seen that the free-kick shot event detection model that combines soccer rules and deep learning effectively improves the performance of free-kick shot event detection and provides a basic reference for further research on the detection of other events in soccer videos.

The paper focused on the study of detrital yellow layer from sediment profile D3 of DaQaidam Salt Lake, based on AMS 14C chronologic and mineralogical methods, aimed to discuss the dust storm events recorded by detrital yellow layer of DaQaidam region since Late Glacial period. The results showed that the grain-size distribution curve displayed double kurtosis, the main peak (10 µm in average) reflected the fine-grained fractions into lake by fluviation, and the secondary peak (> 32 µm) reflected the coarse silt fractions into lake by wind effect and fluviation. Finally, the coarse-grained fractions (> 64 µm) from sediment profile D3 probably recorded the dust storm events since Late Glacial period.

Using an X-ray Fluorescence (XRF) core scanner with nondestructive and successive, the chemistry features of lacustrine sediment can be measured directly. This method of XRF core scanner measurements has been widely applied to core sediment analysis but uncertain of the precision and accuracy. Comparison of intensities obtained by XRF core scanning and the concentration measured by conventional X-ray Fluorescence, indicates effects of physical properties varied from different elements on elemental intensities in the lacustrine sediments of a core from Lake Hurleg in the northeastern Tibetan Plateau. Correlation among elements Ti and Sr between the two measurement methods of the XRF and the conventional XRF is high. Using the intensity of Cl as an indicator of water content, the element intensities of Ti and Sr in the core samples is corrected. But the correlation coefficients of Ti and Sr is litter raised. The results show that XRF core scanning is a very useful tool for measuring element concentration in sediments particularly for high intensities elements.

Little attention has been paid to the study of sedimentary records from DaQaidam Salt Lake regarding brine evolution and hydro-climate change. Here we report the results of AMS 14C dating and mineralogical investigation on a sediment profile D4 from the south central of perennial saline water-body. The result revealed that pinnoite deposit in the lakebed was formed in early Holocene, the lake underwent three stages of lake-level and brine evolution in Holocene period, which were determined mainly by precipitation-controlled catchment inflow. The mineralogical records also depicted fluctuations in lake level and hydrochemistry on decadal and centennial scales.

We target the problem named unsupervised domain adaptive semantic segmentation. A key in this campaign consists in reducing the domain shift, so that a classifier based on labeled data from one domain can generalize well to other domains. With the advancement of adversarial learning method, recent works prefer the strategy of aligning the marginal distribution in the feature spaces for minimizing the domain discrepancy. However, based on the observance in experiments, only focusing on aligning global marginal distribution but ignoring the local joint distribution alignment fails to be the optimal choice. Other than that, the noisy factors existing in the feature spaces, which are not relevant to the target task, entangle with the domain invariant factors improperly and make the domain distribution alignment more difficult. To address those problems, we introduce two new modules, Significance-aware Information Bottleneck (SIB) and Category-level alignment (CLA), to construct a purified embedding-based category-level adversarial network. As the name suggests, our designed network, CLAN, can not only disentangle the noisy factors and suppress their influences for target tasks but also utilize those purified features to conduct a more delicate level domain calibration, i.e., global marginal distribution and local joint distribution alignment simultaneously. In three domain adaptation tasks, i.e., GTA5 $\rightarrow$ → Cityscapes, SYNTHIA $\rightarrow$ → Cityscapes and Cross Season, we validate that our proposed method matches the state of the art in segmentation accuracy. [ABSTRACT FROM AUTHOR]

Aiming at the problem of spoofing flood defense in the control layer of SDN, a controller defense mechanism (CDM)was proposed, including a flood detection mechanism based on key features multi-classification and a flood mitigation mechanism based on SAVI. The flood feature analysis module of the control layer was designed for flood detection, and boosting algorithm was used to overlay each feature weak classifier to form an enhanced classifier, which can achieve more accurate classification spoofing flooding attack effect by continuously reducing the residual in the calculation. In CDM, a flood mitigation mechanism based on SAVI was deployed to realize flood mitigation, which performed flood packet path filtering based on binding-verification mode, and updated the flood features of access layer switches with dynamic polling mode to reduce redundant model update load. The experimental results show that the proposed method has the characteristics of low overhead and high precision. CDM effectively increases the security of the control layer, and reduces the time of host classification of spoofing flood attack and the CPU consumption of corresponding controller. [ABSTRACT FROM AUTHOR]

With the rapid development of data-driven and bandwidth-intensive applications in the Software Defined Networking (SDN) northbound interface, big data stream is dynamically generated with high growth rates in SDN-based data center networks. However, a significant issue faced in big data stream communication is how to verify its authenticity in an untrusted environment. The big data stream traffic has the characteristics of security sensitivity, data size randomness, and latency sensitivity, putting high strain on the SDN-based communication system during larger spoofing events in it. In addition, the SDN controller may be overloaded under big data stream verification conditions on account of the fast increase of bandwidth-intensive applications and quick response requirements. To solve these problems, we propose a two-phase adaptive authenticated model (TAAM) by introducing source address validation implementation- (SAVI-) based IP source address verification. The model realizes real-time data stream address validation and dynamically reduces the redundant verification process. A traffic adaptive SAVI that utilizes a robust localization method followed by the Sequential Probability Ratio Test (SPRT) has been proposed to ensure differentiated executions of the big data stream packets forwarding and the spoofing packets discarding. The TAAM model could filter out the unmatched packets with better packet forwarding efficiency and fundamental security characteristics. The experimental results demonstrate that spoofing attacks under big data streams can be directly mitigated by it. Compared with the latest methods, TAAM can achieve desirable network performance in terms of transmission quality, security guarantee, and response time. It drops 97% of the spoofing attack packets while consuming only 9% of the controller CPU utilization on average. [ABSTRACT FROM AUTHOR]

The annotation of significant events within soccer videos is a fundamental step in content-based soccer video retrieval. In this paper, we propose a soccer video annotation approach based on semantic matching with coarse time constraints, where video events and external text information (match reports) are synchronized using their semantic correspondence in the temporal sequence. Unlike the state-of-the-art soccer video analysis methods that assume that the time of an event’s occurrence is given precisely by the external text information, this paper considers the problem of annotating soccer videos using match reports with coarse-grained time information. The contributions of the proposed approach are: 1) we propose a more generalized approach that synchronizes video events with text descriptions using high-level semantics with coarse time constraints, rather than assuming that the timestamp is given exactly in the text description; 2) the detection of event boundaries is improved by attack–defense transition analysis (ADTA); 3) a robust and fast center circle detection algorithm is proposed for the classification of soccer field zones and ADTA; and 4) unlike conventional audio-based whistle detection, we propose a novel Hough-transform-based algorithm from the perspective of image processing. This allows the game start time to be detected, and further helps the synchronization of video and text events. The experimental results conducted on a large number of soccer videos validate the effectiveness of the proposed approach. [ABSTRACT FROM PUBLISHER]

Chronology is a necessary component of paleoclimatology. Radiocarbon dating plays a central role in determining the ages of geological samples younger than ca. 50 ka BP. However, there are many limitations for its application, including radiocarbon reservoir effects, which may cause incorrect chronology in many lakes. Here we demonstrate temporal changes in the radiocarbon reservoir age of Lake Xingyun, Southwestern China, where radiocarbon ages based on bulk organic matter have been reported in previous studies. Our new radiocarbon ages, determined from terrestrial plant macrofossils suggest that the radiocarbon reservoir age changed from 960 to 2200 years during the last 8500 cal a BP years. These changes to the reservoir effect were associated with inputs from either pre-aged organic carbon or 14C-depleted hard water in Lake Xingyun caused by hydrological change in the lake system. The radiocarbon reservoir age may in return be a good indicator for the carbon source in lake ecosystems and depositional environment. [ABSTRACT FROM AUTHOR]

ABSTRACT We present a pollen-based precipitation reconstruction and multi-proxy records from a 485-cm-long sequence from a sediment core from Xingyun Lake, Yunnan Plateau, south-west China, which depicts the evolution of the Indian Summer Monsoon (ISM) during the last 8500 years. Pollen and other palaeoenvironmental records document several stages of vegetation history and climate change. The warmest and wettest climate in the Xingyun Lake catchment occurred before 5500 cal a BP, and subsequently the climate became gradually drier. After 2000 cal a BP the regional environmental conditions became unstable, and a wet Medieval Warm Period is probably recorded. Our reconstruction of the ISM is similar to that portrayed by Holocene speleothem δ18O records from southern China, but is distinctly different from the East Asian Summer Monsoon (EASM) evolution, which features a mid-Holocene maximum. Our results support the hypothesis that the ISM and EASM evolved asynchronously during the Holocene, and imply that the Chinese speleothem δ18O records from southern China may principally reflect changes in moisture source from the Indian monsoon domain, and thus record the history of the ISM rather than the EASM. [ABSTRACT FROM AUTHOR]

In this research, we propose Projected Residual Vector Quantization (PRVQ) to deal with the problem of large-scale approximate nearest neighbor (ANN) search in a high-dimensional space. A lot of quantization-based ANN search algorithms have been proposed in the past few years. However, most of the existing methods discard the projection errors generated in the dimension reduction process, which inevitably decreases the search accuracy. In view of that, the authors propose a method of projected residual vector quantization for ANN search that considers the projection errors in the quantization process. They also design three simple and effective optimization strategies to improve the performance of the PRVQ algorithm. The authors have integrated the proposed PRVQ algorithm into a mobile landmark recognition system to prove its effectiveness. [ABSTRACT FROM PUBLISHER]

Mobile Visual Location Recognition (MVLR) has attracted a lot of researchers' attention in the past few years. Existing MVLR applications commonly use Query-by-Example (QBE) based image retrieval principle to fulfill the location recognition task. However, the QBE framework is not reliable enough due to the variations in the capture conditions and viewpoint changes between the query image and the database images. To solve the above problem, we make following contributions to the design of a panorama based on-device MVLR system. Firstly, we design a heading (from digital compass) aware BOF (Bag-of-features) model to generate the descriptors of panoramic images. Our approach fully considers the characteristics of the panoramic images and can facilitate the panorama based on-device MVLR to a large degree. Secondly, to search high dimensional visual descriptors directly on mobile devices, we propose an effective bilinear compressed sensing based encoding method. While being fast and accurate enough for on-device implementation, our algorithm can also reduce the memory usage of projection matrix significantly. Thirdly, we also release a panoramas database as well as a set of test panoramic quires which can be used as a new benchmark to facilitate further research in the area. Experimental results prove the effectiveness of the proposed methods for on-device MVLR applications. [ABSTRACT FROM AUTHOR]

Existing mobile visual location recognition (MVLR) applications typically rely on bag-of-features (BOF) representation, which shows superior performance in retrieval accuracy. However, although the BOF framework is promising, it is not compact enough for on-device MVLR. The authors have made two contributions to the design of a BOF-based on-device MVLR system. First, to generate BOF descriptors, they propose a memory-efficient approximate nearest-neighbor search algorithm by combining residual vector quantization (RVQ) and tree-structured RVQ (TSRVQ). Second, they implemented a GPS-based and heading-aware RankBoost algorithm to reduce the dimensionality of the BOF descriptors. The authors evaluate the effectiveness of the proposed algorithms on an HTC mobile phone. Their work applies to on-device MVLR in city-scale workspaces. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of city scale on-device mobile visual location recognition by fusing the inertial sensors and computer vision techniques. The main contributions are as follows: Firstly, we design an efficient vector quantization strategy by combining the Transform Coding (TC) and Residual Vector Quantization (RVQ). Our method can compress a visual descriptor into only several bytes while providing reasonable searching accuracy, which makes the managing of city scale image database directly on mobile devices come true. Secondly, we integrate the information from inertial sensors into the Vector of Locally Aggregated Descriptors (VLAD) generation and image similarity evaluation processes. Our method is not only fast enough for on-device implementation, but it also can improve the location recognition accuracy obviously. Thirdly, we also release a set of 1.295 million geo-tagged street view images with the information from inertial sensors, as well as a difficult set of query images. These resources can be used as a new benchmark to facilitate further research in the area. Experimental results prove the validity of the proposed methods for on-device mobile visual location recognition applications. [ABSTRACT FROM PUBLISHER]

Abstract: Qaidam Basin is a hyperarid inland basin with an area of 121×103 km2 located on the northern Tibetan Plateau. Today, one fourth of the basin is covered by playas and hypersaline lakes. Nearly 80% of brine lithium found in China is contained in four salt lakes: Bieletan (BLT), DongTaijinaier (DT), XiTaijinaier (XT) and Yiliping (YLP). In the past decade, great attention was paid to improving the technology for the extraction of lithium from the brine deposits, but studies on origin and mode of formation of the brine deposits remained limited. Our recent investigations found that: (1) ~748.8t of lithium was transported annually into the lower catchment of the four salt lakes via the Hongshui–Nalinggele River (H–N River), the largest river draining into the Qaidam Basin, (2) Li+-rich brines are formed only in salt lakes associated with inflowing rivers with Li+ concentrations greater than 0.4mg/L, and (3) the water Li+ concentration is positively correlated with both the inflowing river and the associated subsurface brine, including saline lakes with low lithium concentrations. These findings clearly indicate that long-term input of Li+ from the H–N River controls the formation of lithium brine deposits. Here we determine that the source of the lithium is from hydrothermal fields where two active faults converge in the upper reach of the Hongshui River. The hydrothermal fields are associated with a magmatic heat source, as suggested by the high Li+ and As3+ content water from geysers. Based on the assumption of a constant rate of lithium influx, we estimate that the total reserves of lithium were likely formed since the postglacial period. Our data indicate that lithium reserves in each of the four salt lakes depend on the influx of Li+-bearing water from the H–N River. The data also suggest that during the progradation of the alluvial Fan I, the H–N River drained mostly into the BLT salt lake until the Taijinaier River shifted watercourse to the north and began to feed the salt lakes of the DT, XT and YLP, alongside with the Fan II progradation. The inference is consistent with stratigraphic evidence from the sediment cores of the four salt lakes. One of the major findings of our work is the importance of the contrasting hydroclimatic conditions between the high mountains containing ice caps and the terminal salt lakes. The greater than 4000m of relief in the watershed enables a massive amount of ions, such as K+, to be weathered and transported together with detrital material from the huge, relatively wet alpine regions to the hyperarid terminal basins, where intense evaporation rapidly enriches the lake water, resulting in evaporite deposition and associated K+- and Li+-rich brine deposits. [Copyright &y& Elsevier]

Stream programming model has been productively applied to a number of important application domains. Software pipelining is an important code scheduling technique for stream programs. However, the multicore evolution has presented a new dimension of challenges: that is how to orchestrate the best software pipelining schedule in the face of resource constrained architectures (e.g., number of cores, available memory, and bandwidth)? In this paper, we proposed a new solution methodology to address the problem above. Our main contributions include the following. A unified Integer Linear Programming (ILP) formulation has been proposed that combines the requirement of both rate-optimal software pipelining and the minimization of intercore communication overhead. Next, an extended formulation has been proposed to formulate the schedule under memory size constrained systems. It orchestrates the rate-optimal software pipelining execution for stream programs with strict memory, processor cores, and communication constraints. A solution testbed has been implemented for the proposed problem formulations. This has been realized by extending the Brook programming environment with our software pipelining support—named DFBrook. An experimental study has been conducted to verify the effectiveness of the proposed solutions. [ABSTRACT FROM PUBLISHER]

Deep neural networks (DNNs) have shown a great achievement in acoustic modeling for speech recognition task. Of these networks, convolutional neural network (CNN) is an effective network for representing the local properties of the speech formants. However, CNN is not suitable for modeling the long-term context dependencies between speech signal frames. Recently, the recurrent neural networks (RNNs) have shown great abilities for modeling long-term context dependencies. However, the performance of RNNs is not good for low-resource speech recognition tasks, and is even worse than the conventional feed-forward neural networks. Moreover, these networks often overfit severely on the training corpus in the low-resource speech recognition tasks. This paper presents the results of our contributions to combine CNN and conventional RNN with gate, highway, and residual networks to reduce the above problems. The optimal neural network structures and training strategies for the proposed neural network models are explored. Experiments were conducted on the Amharic and Chaha datasets, as well as on the limited language packages (10-h) of the benchmark datasets released under the Intelligence Advanced Research Projects Activity (IARPA) Babel Program. The proposed neural network models achieve 0.1–42.79% relative performance improvements over their corresponding feed-forward DNN, CNN, bidirectional RNN (BRNN), or bidirectional gated recurrent unit (BGRU) baselines across six language collections. These approaches are promising candidates for developing better performance acoustic models for low-resource speech recognition tasks. [ABSTRACT FROM AUTHOR]