In this paper, we have developed a new scheme for achieving multilevel annotations of large-scale images automatically. To achieve more sufficient representation various visual properties the images, both global features and local are extracted image content representation. tackle problem huge intraconcept diversity, multiple types kernels integrated to characterize diverse similarity relationships between precisely, kernel learning algorithm is SVM classifier training. address interconcept...

Automatic image annotation is a promising solution to enable semantic retrieval via keywords. In this paper, we propose multi-level approach annotate the semantics of <b><i>natural scenes</i></b> by using both dominant components (salient objects) and relevant concepts. To achieve automatic at content level, use salient objects as for representation feature extraction. support concept novel classification technique developed map images into most addition, Support Vector Machine (SVM)...

In this paper, we have proposed a novel framework for mining multilevel image semantics via hierarchical classification. To bridge the semantic gap more successfully, salient objects are used to characterize intermediate effectively. The defined as connected regions that capture dominant visual properties linked corresponding physical in an image. achieve reliable and tractable concept learning high-dimensional feature space, algorithm called <i xmlns:mml="http://www.w3.org/1998/Math/MathML"...

It has become a consensus to improve the measurement accuracy of microelectromechanical systems (MEMS) inertial units (IMUs) through an IMU array composed multiple low-cost IMUs. To meet requirements navigation MEMS array, it can be divided into two steps: first step is calibrate single in and second isomorphically fuse gyroscopes accelerometers This article discusses innovative calibration fusion technology based on accompanying test method weighted technology, which optimizes steps array....

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Digital video now plays an important role in medical education, health care, telemedicine and other applications. Several content-based retrieval (CBVR) systems have been proposed the past, but they still suffer from following challenging problems: semantic gap, concept modeling, classification, concept-oriented database indexing access. In this paper, we propose a novel framework to make some advances toward final goal solve these problems. Specifically, includes: 1) semantic-sensitive...

In this paper, a hierarchical classification framework has been proposed for bridging the semantic gap effectively and achieving multi-level image annotation automatically. First, between low-level computable visual features users' real information needs is partitioned into four smaller gaps, multiple approachesallare to bridge these gaps more effectively. To learn reliable contextual relationships atomic concepts co-appearances of salient objects, multi-modal boosting algorithm proposed....

Multi-level annotation of images is a promising solution to enable more effective semantic image retrieval by using various keywords at different levels. In this paper, we propose multi-level approach annotate the semantics natural scenes both dominant components and relevant concepts. contrast well-known image-based region-based approaches, use salient objects as achieve automatic content level. By for representation, novel classification technique developed concept To detect automatically,...

Browsing and retrieving images from large image collections are becoming common important activities. Semantic analysis techniques, which automatically detect high level semantic contents of for annotation, promising solutions toward this problem. However, few efforts have been made to convey the annotation results users in an intuitive manner enable effective browsing retrieval. There is also a lack methods monitor evaluate automatic algorithms due dimensional nature data, features,...

In this paper, we have developed a novel framework called <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">JustClick</i> to enable personalized image recommendation via exploratory search from large-scale collections of Flickr images. First, topic network is automatically generated summarize images at semantic level. Hyperbolic visualization further used interactive navigation and exploration the network, so that users can gain insights first...

Cloud detection is one of the key technologies in field remote sensing. Although extensive deep learning-based cloud methods achieve good performance, their results confusing areas such as boundaries and thin clouds are often not satisfactory due to potential inter-class similarity intra-class inconsistency objects. To this end, we propose a Boundary-Aware Bilateral Fusion network (BABFNet), which effectively enhances by introducing boundary prediction branch an auxiliary. avoid loss...

Most existing content-based video retrieval (CBVR) systems are now amenable to support automatic low-level feature extraction, but they still have limited effectiveness from a user's perspective because of the semantic gap. Automatic concept detection via classification is one promising solution bridge To speed up SVM classifier training in high-dimensional heterogeneous space, novel multimodal boosting algorithm proposed by incorporating hierarchy and reduce both cost size samples...

In this paper, we have developed a novel scheme to incorporate topic network and representativeness-based sampling for achieving semantic visual summarization visualization of large-scale collections Flickr images. First, is automatically generated summarizing visualizing images at level, so that users can select more suitable keywords precise query formulation. Second, the diverse similarities between semantically-similar are characterized precisely by using mixture-of-kernels image...

For online medical education purposes, we have developed a novel scheme to incorporate the results of semantic video classification select most representative shots for generating concept-oriented summarization and skimming surgery videos . First, salient objects are used as patterns feature extraction achieve good representation intermediate semantics. The defined compounds that can be characterize significant perceptual properties corresponding real world physical in video, thus...

To support privacy-preserving video sharing, we have proposed a novel framework that is able to protect the content privacy at individual clip level and prevent statistical inferences from collections. level, developed an effective algorithm automatically detect privacy-sensitive objects events. collections, distributed for classifier training, which significantly reduce costs of data transmission reliably limit breaches by determining optimal size blurred test samples validation. Our...

In this paper, we have developed a novel visualization framework to enable more effective visual analysis of large-scale news videos, where keyframes and keywords are automatically extracted from video clips visually represented according their interestingness measurement help audiences rind stories interest at first glance. A computational approach is also quantify the clips. Our experimental results shown that our techniques for intelligent capacity videos. system very useful security...

The keyword-based Google images search engine is now becoming very popular for online image search. Unfortunately, only the text terms that are explicitly or implicitly linked with used indexing but associated may not have exact correspondence underlying semantics, thus return large amounts of junk which irrelevant to given queries. Based on this observation, we developed an interactive approach filter out from results and our consists following major components. a) A kernel-based clustering...