Search

Cost aggregation is a highly important process in image matching tasks, which aims to disambiguate the noisy matching scores. Existing methods generally tackle this by hand-crafted or CNN-based methods, which either lack robustness to severe deformations or inherit the limitation of CNNs that fail to discriminate incorrect matches due to limited receptive fields and inadaptability. In this paper, we introduce Cost Aggregation with Transformers (CATs) to tackle this by exploring global consensus among initial correlation map with the help of some architectural designs that allow us to fully enjoy global receptive fields of self-attention mechanism. Also, to alleviate some of the limitations that CATs may face, i.e., high computational costs induced by the use of a standard transformer that its complexity grows with the size of spatial and feature dimensions, which restrict its applicability only at limited resolution and result in rather limited performance, we propose CATs++, an extension of CATs. Our proposed methods outperform the previous state-of-the-art methods by large margins, setting a new state-of-the-art for all the benchmarks, including PF-WILLOW, PF-PASCAL, and SPair-71k. We further provide extensive ablation studies and analyses., Accepted to TPAMI. Project page:https://ku-cvlab.github.io/CATs-PlusPlus-Project-Page/ arXiv admin note: text overlap with arXiv:2106.02520

Human rhinovirus (HRV) must be viewed as a significant etiological agent of acute low respiratory tract infections (ALRIs) in infants and young children. The present study has been carried out to investigate the prevalence of recently identified respiratory viruses and determine the phylogenetic composition of HRV strains. In total, 923 nasopharyngeal aspirates (NPAs) collected from hospitalized patients with ALRIs between January and December 2010 (453 females and 470 males) at five referral hospitals in Seoul were tested for respiratory viruses. Viruses were detected by RT-PCR (Reverse transcription polymerase chain reaction) in 501 (54.3%) of the 923 positive samples obtained. Overall, HRV was detected in 7.9%, ADV in 10.2%, IFV in 21.9%, RSV in 3.6%, hMPV in 2.7%, HCoV in 2.5%, HEV in 1.6%, HBoV in 1.6%, and PIV in 2.4% of the samples. HRV infections occurred throughout the year, with peaks in January, May, July, and November of 2010. Phylogenetic analyses using the VP4/VP2 cording region showed that among the 27 isolates with HRV, 14 (52%) were infected with species A, 2 (7%) were infected with species B, and 11 (41%) were infected with the strains from HRV-C. HRV-C sequences are genetically distinct, sharing only 32% to 33% of their amino acids with HRV-A and HRV-B, while retaining 46% to 100% identity with each other. HRV-A, B, and C were co-circulating in children hospitalized with ALRIs in Korea in 2010. These findings indicated that HRV-C was the important causative agent of HRV associated with ALRIs.

The present study has determined the detection rate of norovirus (NoV) with acute gastroenteritis (AGE) in hospitalized children and describes the molecular epidemiology of NoV circulating in Seoul, Korea. Six hundred and eighty-three (9.8%) of samples were positive for NoV. Of these, the NoV GII genogroup was the most commonly found, with a prevalence of 96.2% (683 of 710). Only 27 samples were positive for the NoV GI genogroup. Ten kinds of GI genotype (GI/1, GI/2, GI/3, GI/4, GI/5, GI/6, GI/7, GI/9, GI/12, and GI/13) and eight kinds of GII genotype (GII/2, GII/3, GII/4, GII/8, GII/14, GII/15, GII/16, and GII/17) were identified in children with AGE during the years 2008-2011.

In total, 4,806 stool samples collected from 2003 to 2009 were tested for norovirus (NVs) RNA. A positive polymerase chain reaction (PCR) result was obtained from a total of 628 (13%) of the samples. Molecular analysis of strains from 64 outbreaks showed that there were up to 20 genotypes involved in the outbreaks. The dominant genotype throughout the study period was GII-4. In this study, NVs are an important cause of gastroenteritis in Seoul, Korea. By this molecular investigation, genotypic distribution associated with NVs infections will be used for control and prevention of NVs related diseases.

The present study has determined the detection rate of norovirus (NoV) with acute gastroenteritis (AGE) in hospitalized children and describes the molecular epidemiology of NoV circulating in Seoul, Korea. Six hundred and eighty-three (9.8%) of samples were positive for NoV. Of these, the NoV GII genogroup was the most commonly found, with a prevalence of 96.2% (683 of 710). Only 27 samples were positive for the NoV GI genogroup. Ten kinds of GI genotype (GI/1, GI/2, GI/3, GI/4, GI/5, GI/6, GI/7, GI/9, GI/12, and GI/13) and eight kinds of GII genotype (GII/2, GII/3, GII/4, GII/8, GII/14, GII/15, GII/16, and GII/17) were identified in children with AGE during the years 2008-2011.