Wireless capsule endoscopy (WCE) is a minimally invasive procedure enabling visual examination of the entire gastrointestinal tract. It is performed by a swallow able camera-equipped capsule wirelessly transmitting co...
详细信息
Wireless capsule endoscopy (WCE) is a minimally invasive procedure enabling visual examination of the entire gastrointestinal tract. It is performed by a swallow able camera-equipped capsule wirelessly transmitting color video frames. Reading the resulting video is a challenging process for the endoscopists. It requires a considerable amount of time, usually hours of undivided attention and experience to detect and correctly interpret findings. In this paper we propose a novel method for automatic construction of panoramic visual summaries of WCE video frames enabling efficient reading of the video, while providing a broader field of view for the endoscopists. It involves extraction of visual features from each frame, and frame registration. Consecutive frames are geometrically transformed and stitched together to form new sequences of panoramic frames, each of which summarizes the information contained in multiple original ones. The proposed method has been integrated with Java Video Analysis (JVA), a novel software development framework for video applications. Its experimental evaluation on real, publicly available WCE video clips show that its application can result in a reduction of 84.5%of the video frames to be examined by the endoscopes.
In Long Term Evolution (LTE) networks, the Multimedia Broadcast/Multicast Services over Single Frequency Network (MBSFN) operation enables the delivery of the same multimedia content to a set or all the users in a cel...
详细信息
In Long Term Evolution (LTE) networks, the Multimedia Broadcast/Multicast Services over Single Frequency Network (MBSFN) operation enables the delivery of the same multimedia content to a set or all the users in a cell, leading in this way to efficient utilization of the network resources. This is achieved through a time-synchronized common waveform that is transmitted from multiple cells for a given duration, which enables over-the-air combining, and therefore significantly improves the spectral efficiency. However, the performance of MBSFN is directly connected to the Modulation and Coding Scheme (MCS) and the transmission mode that will be selected for the delivery of the multimedia data. This paper proposes three different approaches for the selection of the MCS during MBSFN transmissions. Main target is to determine the most efficient approach depending on the target that the operator may set, i.e. the approach that either maximizes the spectral efficiency or achieves a target spectral efficiency for various user distributions. The performance of the approaches is evaluated for scenarios that include users' mobility and for three different transmission modes in order to also examine the effectiveness of Multiple Input Multiple Output (MIMO) techniques.
The ShARe/CLEF eHealth Evaluation Lab (SHEL) organized a challenge on natural language processing (NLP) and information retrieval (IR) in the medical domain in 2013. The first task of the 2013 ShARe/CLEF challenge was...
详细信息
The ShARe/CLEF eHealth Evaluation Lab (SHEL) organized a challenge on natural language processing (NLP) and information retrieval (IR) in the medical domain in 2013. The first task of the 2013 ShARe/CLEF challenge was to extract disorder mention spans and their associated UMLS (Unified Medical Language System) concept unique identifiers (CUIs). We participated in Task 1 and developed a clinical disorder recognition and encoding system. The proposed system consists of two components: a machine learning-based approach to recognize disorder entities and a vector space model-based method to encode disorders to UMLS CUIs. The challenge organizers manually annotated disorder entities and corresponding UMLS CUIs in 298 clinical notes, of which 199 notes were used for training and 99 were for testing. Evaluation on the test data set showed that our system achieved the best F-measure of 0.750 for entity recognition (ranked first) and the highest F-measure of 0.514 for UMLS CUI encoding (ranked third), indicating the promise of the proposed approaches.
Current trends in Future Internet, 3D Internet and Internet of Things are highlighting the possibility that virtual spaces will be available in the near future for the creation of virtual/physical collaboration spaces...
详细信息
In this paper we present a tool for gathering touchscreen data for usability analysis. Touchscreens are becoming the most popular input/output devices mainly because of increasing number of smartphones and tablets. As...
详细信息
Overlay networks are logical networks running on the highest level of the OSI stack: they are applicative networks used by millions of users everyday. In many scenarios, it would be desirable for peers belonging to ov...
详细信息
ISBN:
(纸本)9781450316569
Overlay networks are logical networks running on the highest level of the OSI stack: they are applicative networks used by millions of users everyday. In many scenarios, it would be desirable for peers belonging to overlays running different protocols to communicate with each other and exchange certain information. However, due to differences in their respective protocols, this communication is often difficult or even impossible to be achieved efficiently, even if the overlays are sharing common objectives and functionalities. In this paper, we address this problem by presenting a new overlay protocol, called OGP (Overlay Gateway Protocol), allowing different existing networks to route messages between each other in a backward-compatible fashion, by making use of specialized peers joined together into a super-overlay. Experimental results on a large scale Grid5000 infrastructure show that having only a small number of nodes running the OGP protocol is sufficient for achieving efficient routing between heterogeneous overlay networks. Copyright 2013 ACM.
Spatial selections are a ubiquitous concept in visualization. By localizing particular features, they can be analyzed and compared in different views. However, the semantics of such selections are often dependent on o...
详细信息
ISBN:
(纸本)9781450324809
Spatial selections are a ubiquitous concept in visualization. By localizing particular features, they can be analyzed and compared in different views. However, the semantics of such selections are often dependent on other parameter settings and it can be difficult to reconstruct them without additional information. In this paper, we present the concept of contextual snapshots as an effective means for managing spatial selections in visualized data. The selections are automatically associated with the context in which they have been created. Contextual snapshots can be also used as the basis for interactive integrated and linked views, which enable in-place investigation and comparison of multiple visual representations of data. Our approach is implemented as a flexible toolkit with well-defined interfaces for integration into existing systems. We demonstrate the power and generality of our techniques by applying them to several distinct scenarios such as the visualization of simulation data and the analysis of historical documents.
Professional search in specific domains (e.g. patent, medical, scientific literature, media) usually needs an exploratory type of search which is characterized more often, in comparison to fact finding and question an...
详细信息
Different logotypes represent significant cues for video annotations. A combination of temporal and spatial segmentation methods can be used for logo extraction from various video contents. To achieve this segmentatio...
详细信息
暂无评论