Temporal action localization is an important yet challenging problem. Given a long, untrimmed video consisting of multiple action instances and complex background contents, we need not only to recognize their action categories, but also to localize the start time and end time of each instance. Many state-of-the-art systems use segmentlevel classifiers to select and rank proposal segments of predetermined boundaries. However, a desirable model should move beyond segment-level and make dense predictions at a fine granularity in time to determine precise temporal boundaries. To this end, we design a novel Convolutional-De-Convolutional (CDC) network that places CDC filters on top of 3D ConvNets, which have been shown to be effective for abstracting action semantics but reduce the temporal length of the input data. The proposed CDC filter performs the required temporal upsampling and spatial downsampling operations simultaneously to predict actions at the frame-level granularity. It is unique in jointly modeling action semantics in space-time and fine-grained temporal dynamics. We train the CDC network in an end-toend manner efficiently. Our model not only achieves superior performance in detecting actions in every frame, but also significantly boosts the precision of localizing temporal boundaries. Finally, the CDC network demonstrates a very high efficiency with the ability to process 500 frames per second on a single GPU server. Source code and trained models are available online at https://bitbucket. org/columbiadvmm/cdc.

Human-induced land use changes are causing extensive habitat fragmentation. As a result, many species are not able to shift their ranges in response to climate change and will likely need to adapt in situ to changing climate conditions. Consequently, a prudent strategy to maintain the ability of populations to adapt is to focus conservation efforts on areas where levels of intraspecific variation are high. By doing so, the potential for an evolutionary response to environmental change is maximized. Here, we use modeling approaches in conjunction with environmental variables to model species distributions and patterns of genetic and morphological variation in seven Ecuadorian amphibian, bird, and mammal species. We then used reserve selection software to prioritize areas for conservation based on intraspecific variation or species-level diversity. Reserves selected using species richness and complementarity showed little overlap with those based on genetic and morphological variation. Priority areas for intraspecific variation were mainly located along the slopes of the Andes and were largely concordant among species, but were not well represented in existing reserves. Our results imply that in order to maximize representation of intraspecific variation in reserves, genetic and morphological variation should be included in conservation prioritization.

The results of digital image analysis suggest that although increased FAF may have a modest PPV for new GA development, the relative PPV is generally no greater than chance. Similarly, the relative NPV demonstrates negligible difference from chance and is also lower than the relative NPV of proximity. This suggests that increased FAF, though a disease manifestation, is not a strong risk factor for development or extension of GA.

No abstract

Group B coxsackieviral (CVB) infection commonly causes viral myocarditis. Mice are protected from CVB3 myocarditis by gene-targeted knockout of p56(Lck)(Lck), the Src family kinase (Src) essential for T cell activation. Extracellular signal-regulated kinase 1 and 2 (ERK-1/2) can influence cell function downstream of Lck. Using T cell lines and neonatal cardiac myocytes we investigated the role of ERK-1/2 in CVB3 infection. In Jurkat T cells ERK-1/2 is rapidly activated by CVB3; but, this response is absent in Lck-negative JCaM T cells. Inhibition of ERK-1/2 with UO126 reduced CVB3 titers in Jurkat cells, but not in JCaM cells. In cardiac myocytes CVB3 activation of ERK-1/2 is blocked by the Src inhibitor PP2. In addition, viral production in myocytes is decreased by Src or ERK-1/2 inhibition. In vitro, in both immune and myocardial cells, ERK-1/2 is activated by CVB3 downstream of Lck and other Src's and is necessary for efficient CVB3 replication. In vivo, following CVB3 infection, ERK-1/2 activation is evident in the myocardium. ERK-1/2 activation is intense in the hearts of myocarditis-susceptible A/J mice. In contrast, significantly less ERK-1/2 activation is found in the hearts of myocarditis-resistant C57BL/6 mice. Therefore, the ERK-1/2 response to CVB3 infection may contribute to differential host susceptibility to viral myocarditis.

No abstract

No abstract

We aim to tackle a novel task in action detection -Online Detection of Action Start (ODAS) in untrimmed, streaming videos. The goal of ODAS is to detect the start of an action instance, with high categorization accuracy and low detection latency. ODAS is important in many applications such as early alert generation to allow timely security or emergency response. We propose three novel methods to specifically address the challenges in training ODAS models: (1) hard negative samples generation based on Generative Adversarial Network (GAN) to distinguish ambiguous background, (2) explicitly modeling the temporal consistency between data around action start and data succeeding action start, and (3) adaptive sampling strategy to handle the scarcity of training data. We conduct extensive experiments using THUMOS'14 and ActivityNet. We show that our proposed methods lead to significant performance gains and improve the state-of-the-art methods. An ablation study confirms the effectiveness of each proposed method.