ORCID Profile
0000-0003-4726-847X
Current Organisations
University of Reading
,
University of Electronic Science and Technology of China
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: Elsevier BV
Date: 08-2020
Publisher: Elsevier BV
Date: 07-2022
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 07-2022
Publisher: Association for Computing Machinery (ACM)
Date: 04-03-2022
DOI: 10.1145/3503927
Abstract: Fully mining visual cues to aid in content understanding is crucial for video captioning. However, most state-of-the-art video captioning methods are limited to generating captions purely based on straightforward information while ignoring the scenario and context information. To fill the gap, we propose a novel, simple but effective scenario-aware recurrent transformer (SART) model to execute video captioning. Our model contains a “scenario understanding” module to obtain a global perspective across multiple frames, providing a specific scenario to guarantee a goal-directed description. Moreover, for the sake of achieving narrative continuity in the generated paragraph, a unified recurrent transformer is adopted. To demonstrate the effectiveness of our proposed SART, we have conducted comprehensive experiments on various large-scale video description datasets, including ActivityNet, YouCookII, and VideoStory. Additionally, we extend a story-oriented evaluation framework for assessing the quality of the generated caption more precisely. The superior performance has shown that SART has a strong ability to generate correct, deliberative, and narrative coherent video descriptions.
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 08-2023
Publisher: Wiley
Date: 11-12-2022
DOI: 10.1002/QJ.4378
Abstract: While skilful forecasts of heavy rainfall are highly desirable for weather warnings and mitigating impacts, forecasting such events is notoriously difficult, even with the most advanced numerical weather prediction models, due to the strong dependence on convective‐scale processes. The large‐scale circulation, on the other hand, is typically more predictable. Weather patterns (WPs) are a set of circulation types obtained statistically that can be used to characterize regional weather and harness the predictability of the large‐scale circulation. In this work we produce pattern‐conditioned probabilistic rainfall forecasts by projecting the horizontal winds from the Met Office GloSea5 prediction system on to WPs and then using the observed relationship between each WP and rainfall estimated by satellite. The WPs are derived following a novel two‐tier clustering technique: the WPs in the first tier represent planetary‐scale variability, such as El Niño–Southern Oscillation (ENSO), while the WPs in the second tier capture synoptic‐scale variability. We investigate WP predictability as well as the improvement in skill of subseasonal rainfall forecasts gained by this technique. GloSea5 predicts the WP occurrence with skill extending beyond lead day 10. The pattern‐conditioned rainfall forecasts were evaluated against climatological forecasts and model‐simulated rainfall hindcasts. We show that the pattern‐conditioned forecasts are skilful and outperform the model‐simulated rainfall hindcasts for lead times extending to days 10–20, depending on the specific exceedance criteria and region. Spatial aggregation leads to increased levels of skill, but not to a significant extension of the skilful prediction horizon. These results constitute a fundamental step for the development of subseasonal prediction systems for Southeast Asia.
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
Date: 2022
Publisher: Springer Science and Business Media LLC
Date: 06-03-2019
Publisher: Elsevier BV
Date: 2019
Location: United Kingdom of Great Britain and Northern Ireland
Location: United Kingdom of Great Britain and Northern Ireland
Location: China
No related grants have been discovered for Samantha Ferrett.