Based on keypoints extracted as salient image patches, an can be described a "bag of visual words" and this representation has been used in scene classification. The choice dimension, selection, weighting words is crucial to the classification performance but not thoroughly studied previous work. Given analogy between bag-of-words text documents, we apply techniques categorization, including term weighting, stop word removal, feature generate representations that differ words. impact these...

Many multimedia applications can benefit from techniques for adapting existing classifiers to data with different distributions. One example is cross-domain video concept detection which aims adapt across various domains. In this paper, we explore two key problems classifier adaptation: (1) how transform classifier(s) into an effective a new dataset that only has limited number of labeled examples, and (2) select the best adaptation. For first problem, propose Adaptive Support Vector...

Bag-of-features (BoF) deriving from local keypoints has recently appeared promising for object and scene classification. Whether BoF can naturally survive the challenges such as reliability scalability of visual classification, nevertheless, remains uncertain due to various implementation choices. In this paper, we evaluate factors which govern performance BoF. The include choices detector, kernel, vocabulary size weighting scheme. We offer some practical insights in how optimize by choosing...

Based on the local keypoints extracted as salient image patches, an can be described a ¿bag-of-visual-words (BoW)¿ and this representation has appeared promising for object scene classification. The performance of BoW features in semantic concept detection large-scale multimedia databases is subject to various choices. In paper, we conduct comprehensive study choices BoW, including vocabulary size, weighting scheme, stop word removal, feature selection, spatial information, visual bi-gram....

Mobile edge computing (MEC) is proposed as a new paradigm to meet the ever-increasing computation requirements, which caused by rapid growth of Internet Things (IoT) devices. As supplement terrestrial network, satellites can provide communication devices in some harsh environments and natural disasters. Satellite becoming an emerging topic technology. In this paper, game-theoretic approach optimization offloading strategy satellite proposed. The system model for established, considering...

Combining retrieval results from multiple modalities plays a crucial role for video systems, especially automatic systems without any user feedback and query expansion. However, most of current only utilize independent combination or rely on explicit weighting. In this work, we propose using query-class dependent weights within hierarchial mixture-of-expert framework to combine results. We first classify each into one the four predefined categories then aggregate with associated weights,...

Many data mining applications can benefit from adapt- ing existing classifiers to new with shifted distribu- tions. In this paper, we present Adaptive Support Vector Machine (Adapt-SVM) as an efficient model for adapting a SVM classifier trained one dataset where only limited labeled examples are available. By in- troducing regularizer into SVM's objective function, Adapt-SVM aims minimize both the classification error over training examples, and discrepancy between adapted original...

Cooperative jamming for multiuser multiple input output (MIMO) broadcast channel is studied to enhance the physical layer security with help of a friendly jammer. We assume base station transmits independent data streams legitimate users. During transmission, however, there are eavesdroppers antennas that have interests in from station. In order wiretap desired streams, may collude or not, and maximize signal-to-interference-plus-noise ratio (SINR) using receive beamforming. The optimal...

As one of the information industry's future development directions, Internet Things (IoT) has been widely used. In order to reduce pressure on network caused by long distance between processing platform and terminal, edge computing provides a new paradigm for IoT applications. many scenarios, devices are distributed in remote areas or extreme terrain cannot be accessed directly through terrestrial network, data transmission can only achieved via satellite. However, traditional satellites...

We present an efficient system for video search that maximizes the use of human bandwidth, while at same time exploiting machine's ability to learn in real-time from user selected relevant clips. The exploits capability rapidly scanning imagery augmenting it with active learning loop, which attempts always most material based on current information. Two versions interface were evaluated, one variable page sizes and manual paging, other a fixed size automatic paging. Both require absolute...

Labeling faces in news video with their names is an interesting research problem which was previously solved using supervised methods that demand significant user efforts on labeling training data. In this paper, we investigate a more challenging setting of the where there no complete information data labels. Specifically, by exploiting uniqueness face's name, formulate as special multi-instance learning (MIL) problem, namely exclusive MIL or eMIL so it can be tackled model trained partial...

Abstract. Obtaining an accurate cloud-cover state is a challenging task. In the past, traditional two-dimensional red-to-blue band methods have been widely used for cloud detection in total-sky images. By analyzing imaging principle of cameras, green channel has selected to replace 2-D detecting pixels from partly cloudy images this study. The brightness distribution image usually nonuniform, because forward scattering and Mie aerosols, which results increased errors circumsolar near-horizon...

This paper proposes a restricted weighted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${k}$ </tex-math></inline-formula> -nearest neighbor algorithm (RWKNN) specific for indoor environments. The traditional WKNN method determines the locations by calculating difference between current received signal strength (RSS) and fingerprint RSS. However, limitations of positioning include RSS instability spatial...

Naming every individual person appearing in broadcast news videos with names detected from the video transcript leads to better access of content. In this paper, we approach challenging problem a statistical learning method. Two categories information extracted multiple modalities have been explored, namely <i>features</i>, which help distinguish true name person, as well <i>constraints</i>, reveal relationships among different persons. The person-naming is formulated into framework predicts...

Great effort has been made to improve video concept detection and continuous progress reported. With the current evaluation method being confined carefully annotated domains thus quite forgiving, reliability of state-of-the-art classifiers remains in question. Adopting a more rigorous approach, we find that most built using mainstream approach are unreliable because they generalize poorly other than their training domain. Moreover, evidences show SVM-based learn little beyond memorizing...

Linear precoding and cooperative jamming for multiuser broadcast channel is studied to enhance the physical layer security. We consider system where multiple independent data streams are transmitted from base station legitimate users with help of a friendly jammer. It assumed that normalized linear matrix given at station, whereas power allocated each user be determined. The problem jointly design allocation across different goal maximize lower bound secrecy rate, provided minimum...

Many factors influence the reliable assessment of medical students' competencies in clerkships. The purpose this study was to determine how many clerkship competency scores were necessary achieve an acceptable threshold reliability.Clerkship student data collected during 2015-2016 academic year as part school program at University Michigan Medical School. Faculty and residents assigned for third-year core students. Generalizability (G) decision (D) studies conducted using balanced,...

With the continuous development of modern society, rapid expansion human civilization squeezes living space other organisms, and extinction more biological species has sounded alarm for us. Therefore, in order to timely understand changes wild animals resources a specific area, facilitate relevant personnel formulate effective restoration protection measures, this paper proposes Animal recognition network based on deep learning improved YOLOv5: YOLO-Animal. The application artificial...

Temporal consistency is ubiquitous in video data, where temporally adjacent shots usually share similar visual and semantic content.This paper presents a thorough study of temporal defined with respect to concepts query topics using quantitative measures,and discusses its implications analysis retrieval tasks. We further show that,in interactive settings, leads considerable improvement on the performance concept detection data.Speci fically,an active learning method sampling strategy...

Precisely evaluating the systematic error induced by quadratic Zeeman effect is important for developing atom interferometer gravimeters aiming at an accuracy in $\ensuremath{\mu}\mathrm{Gal}$ regime ($1\ensuremath{\mu}\mathrm{Gal}={10}^{\ensuremath{-}8}\mathrm{m}/{\mathrm{s}}^{2}\ensuremath{\approx}{10}^{\ensuremath{-}9}\mathrm{g}$). This paper reports on experimental investigation of Raman spectroscopy-based magnetic field measurements and evaluation gravimetric (GAIN) due to effect. We...