Visual tracking, in essence, deals with nonstationary image streams that change over time. While most existing algorithms are able to track objects well in controlled environments, they usually fail in the presence of significant variation of the object's appearance or surrounding illumination. One reason for such failures is that many algorithms employ fixed appearance models of the target. Such models are trained using only appearance data available before tracking begins, which in practice limits the range of appearances that are modeled, and ignores the large volume of information (such as shape changes or specific lighting conditions) that becomes available during tracking. In this paper, we present a tracking method that incrementally learns a low-dimensional subspace representation, efficiently adapting online to changes in the appearance of the target. The model update, based on incremental algorithms for principal component analysis, includes two important features: a method for correctly updating the sample mean, and a forgetting factor to ensure less modeling power is expended fitting older observations. Both of these features contribute measurably to improving overall tracking performance. Numerous experiments demonstrate the effectiveness of the proposed tracking algorithm in indoor and outdoor environments where the target objects undergo large changes in pose, scale, and illumination.
This paper introduces a video dataset of spatiotemporally localized Atomic Visual Actions (AVA). The AVA dataset densely annotates 80 atomic visual actions in 430 15-minute video clips, where actions are localized in space and time, resulting in 1.58M action labels with multiple labels per person occurring frequently. The key characteristics of our dataset are: (1) the definition of atomic visual actions, rather than composite actions;(2) precise spatio-temporal annotations with possibly multiple annotations for each person; (3) exhaustive annotation of these atomic actions over 15-minute video clips; (4) people temporally linked across consecutive segments; and (5) using movies to gather a varied set of action representations. This departs from existing datasets for spatio-temporal action recognition, which typically provide sparse annotations for composite actions in short video clips.AVA, with its realistic scene and action complexity, exposes the intrinsic difficulty of action recognition. To benchmark this, we present a novel approach for action localization that builds upon the current state-of-the-art methods, and demonstrates better performance on JHMDB and UCF101-24 categories. While setting a new state of the art on existing datasets, the overall results on AVA are low at 15.6% mAP, underscoring the need for developing new approaches for video understanding.
We propose TAL-Net, an improved approach to temporal action localization in video that is inspired by the Faster R-CNN object detection framework. TAL-Net addresses three key shortcomings of existing approaches: (1) we improve receptive field alignment using a multi-scale architecture that can accommodate extreme variation in action durations;(2) we better exploit the temporal context of actions for both proposal generation and action classification by appropriately extending receptive fields; and (3) we explicitly consider multi-stream feature fusion and demonstrate that fusing motion late is important. We achieve state-ofthe-art performance for both action proposal and localization on THUMOS'14 detection benchmark and competitive performance on ActivityNet challenge.
Familial combined hyperlipidemia (FCHL, MIM-144250) is a common, multifactorial and heterogeneous dyslipidemia predisposing to premature coronary artery disease and characterized by elevated plasma triglycerides, cholesterol, or both. We identified a mutant mouse strain, HcB-19/Dem (HcB-19), that shares features with FCHL, including hypertriglyceridemia, hypercholesterolemia, elevated plasma apolipoprotein B and increased secretion of triglyceride-rich lipoproteins. The hyperlipidemia results from spontaneous mutation at a locus, Hyplip1, on distal mouse chromosome 3 in a region syntenic with a 1q21-q23 FCHL locus identified in Finnish, German, Chinese and US families. We fine-mapped Hyplip1 to roughly 160 kb, constructed a BAC contig and sequenced overlapping BACs to identify 13 candidate genes. We found substantially decreased mRNA expression for thioredoxin interacting protein (Txnip). Sequencing of the critical region revealed a Txnip nonsense mutation in HcB-19 that is absent in its normolipidemic parental strains. Txnip encodes a cytoplasmic protein that binds and inhibits thioredoxin, a major regulator of cellular redox state. The mutant mice have decreased CO2 production but increased ketone body synthesis, suggesting that altered redox status down-regulates the citric-acid cycle, sparing fatty acids for triglyceride and ketone body production. These results reveal a new pathway of potential clinical significance that contributes to plasma lipid metabolism.
Differences in sweetener intake among inbred strains of mice are partially determined by allelic variation of the saccharin preference (Sac) locus. Genetic and physical mapping limited a critical genomic interval containing Sac to a 194 kb DNA fragment. Sequencing and annotation of this region identified a gene (Tas1r3) encoding the third member of the T1R family of putative taste receptors, T1R3. Introgression by serial backcrossing of the 194 kb chromosomal fragment containing the Tas1r3 allele from the high-sweetener-preferring C57BL/6ByJ strain onto the genetic background of the low-sweetener-preferring 129P3/J strain rescued its low-sweetener-preference phenotype. Polymorphisms of Tas1r3 that are likely to have functional significance were identified using analysis of genomic sequences and sweetener-preference phenotypes of genealogically distant mouse strains. Tas1r3 has two common haplotypes, consisting of six single nucleotide polymorphisms: one haplotype was found in mouse strains with elevated sweetener preference and the other in strains relatively indifferent to sweeteners. This study provides compelling evidence that Tas1r3 is equivalent to the Sac locus and that the T1R3 receptor responds to sweeteners.
The process of adipogenesis involves a complex program of gene expression that includes down-regulation of the gene encoding Hes-1, a target of the Notch signaling pathway. To determine if Notch signaling affects adipogenesis, we exposed 3T3-L1 preadipocytes to the Notch ligand Jagged1 and found that differentiation was significantly reduced. This effect could be mimicked by constitutive expression of Hes-1. The block was associated with a complete loss of C/EBP␣ and peroxisome proliferator-activated receptor ␥ (PPAR␥) induction and could be overcome by retroviral expression of either C/EBP␣ or PPAR␥2. Surprisingly, small interfering RNA (siRNA)-mediated reduction of Hes-1 mRNA in 3T3-L1 cells also inhibited differentiation, suggesting an additional, obligatory role for Hes-1 in adipogenesis. This role may be related to our observation that both Notch signaling and Hes-1 down-regulate transcription of the gene encoding DLK/Pref-1, a protein known to inhibit differentiation of 3T3-L1 cells. The results presented in this study establish a new target downstream of the Notch-Hes-1 pathway and suggest a dual role for Hes-1 in adipocyte development.
Charcot-Marie-Tooth disease type 1A (CMT1A) is associated with a DNA duplication at chromosome 17p11.2. In view of the point mutation in the gene for peripheral myelin protein pmp-22/gas-3 in Trembler mice, a murine model for CMT1A, we have analysed whether this gene is altered in CMT1A. Here we show that the human homologue of the murine pmp-22 gene is located within the CMT1A DNA duplication, which is a direct repeat and does not interrupt the coding region of PMP-22. Expression of PMP-22 in CMT1A fibroblasts is similar to expression in control fibroblasts. Increased gene dosage or altered PMP-22 expression in the peripheral nervous system are therefore possible mechanisms by which PMP-22 is involved in CMT1A.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.