Cascaded Prediction Network via Segment Tree for Temporal Video Grounding

Author(s):  
Yang Zhao ◽  
Zhou Zhao ◽  
Zhu Zhang ◽  
Zhijie Lin
Keyword(s):  
2008 ◽  
Vol 89 (8) ◽  
pp. 1987-1997 ◽  
Author(s):  
Yang Zou ◽  
Jing Hu ◽  
Zhao-Xiao Wang ◽  
Ding-Ming Wang ◽  
Ming-Hui Li ◽  
...  

To gain further insight into the molecular epidemiology of Hantaan virus (HTNV) in Guizhou, China, rodents were captured in this region in 2004 and 2005. In addition, serum samples were collected from four patients. Ten hantaviruses were isolated successfully in cell culture from four humans, two Apodemus agrarius, three Rattus norvegicus and one Rattus nitidus. The nucleotide sequences for their small (S), medium (M) and partial large (L) segments were determined. Phylogenetic analysis of the S and M segment sequences revealed that all of these isolates belong to the species HTNV, suggesting a spillover of HTNV from A. agrarius to Rattus rats. All available isolates from Guizhou were divided into four distinct groups either in the S segment tree or in the M segment tree. The clustering pattern of these isolates in the S segment tree was not in agreement with that in the M or L segment tree, showing that genetic reassortment between HTNV had occurred naturally. Analysis of the S segment sequences from available HTNV strains indicated that they formed three clades. The first clade, which comprised only viruses from Guizhou, was the outgroup of clades II and III. The viruses in the second clade were found in Guizhou and mainly in the far-east Asian region, including China. However, the viruses in the third clade were found in most areas of China, including Guizhou, in which haemorrhagic fever with renal syndrome (HFRS) is endemic. Our results reveal that the highest genetic diversity of HTNV is in a limited geographical region of Guizhou, and suggest that Guizhou might be a radiation centre of the present form of HTNV.


Author(s):  
Yanyan Xu ◽  
◽  
Xiangyang Xu ◽  
Rui Yu

A disparity optimization algorithm based on an improved guided filter is proposed to smooth the disparity image. A well-known problem to local stereo matching is the low matching accuracy and staircase effect in regions with weak texture and slope. Our disparity optimization method solves this problem and achieve a smooth disparity. First, the initial disparity image is obtained by a local stereo matching algorithm using segment tree. Then, the guided filter is improved by using gradient domain information. Lastly, the improved guided filter is adopted as the disparity optimization method to smooth the disparity image. Experiments conducted on the Middlebury data sets demonstrate that by using the proposed algorithm in this paper, the smoothness of the disparity map in slope regions is improved, and a higher precision of dense disparity is obtained.


1990 ◽  
Vol 11 (9) ◽  
pp. 619-623 ◽  
Author(s):  
Anita G. Pai ◽  
H. Usha ◽  
Arun K. Pujari
Keyword(s):  

2014 ◽  
Vol 602-605 ◽  
pp. 1626-1629
Author(s):  
Chen Zhang ◽  
Yu Quan Chen

In order to identify new words in huge Chinese corpus efficiently, this paper comes up with an algorithm based on ensemble methods. At first we perform Chinese word segmenting with Trie and build segment-tree. Then we select words pattern drawing method, frequency filtering, independent word probability and naive Bayes model to be sub-models of ensemble methods and train them independently. At last we integrate results from different sub-models with a multi-layer model. In experiment, this algorithm is proved to be quite fast as well as product precise and high-coverage results.


2021 ◽  
Vol 12 (5) ◽  
Author(s):  
Josué Ttito ◽  
Renato Marroquín ◽  
Sergio Lifschitz ◽  
Lewis McGibbney ◽  
José Talavera

Key-value stores propose a straightforward yet powerful data model. Data is modeled using key-value pairs where values can be arbitrary objects and written/read using the key associated with it. In addition to their simple interface, such data stores also provide read operations such as full and range scans. However, due to the simplicity of its interface, trying to optimize data accesses becomes challenging. This work aims to enable the shared execution of concurrent range and point queries on key-value stores. Thus, reducing the overall data movement when executing a complete workload. To accomplish this, we analyze different possible data structures and propose our variation of a segment tree, Updatable Interval Tree. Our data structure helps us co-planning and co-executing multiple range queries together and reduces redundant work. This results in executing workloads more efficiently and overall increased throughput, as we show in our evaluation.


Sign in / Sign up

Export Citation Format

Share Document