Remote photoplethysmography (rPPG)-based heart rate (HR) measurement approaches have attracted increasing attention recently. In recent deep-learning-based approaches, the inherent redundancy and noise in video frame contents are urgent issues to be addressed. Some studies utilize spatial–temporal maps (STmaps) to encode the video clips into efficient spatial–temporal representations. Since the subtle physiological signals are vulnerable to external factors such as head movement and unstable illumination, how to construct a better physiological representation is a significant factor affecting the method's performance. In addition, existing STmap-based methods always neglect temporal modeling between video clips, which is very important for stable video HR estimation. To address the above problems, in this article, we propose an effective temporal shift attention network with enhanced STmap as its input (EST-TSANet). We first propose an enhanced STmap (ESTmap), which combines both local and global physiological information to construct a better physiological representation. To establish the temporal relationship between multiple STmaps, we construct a split temporal shift block (STSB) and leverage temporal shift to perform efficient temporal modeling. Besides, considering the characteristics of STmap, a novel plug-and-play module named dimension-specific attention module (DSAM) is designed to optimize the information learning specifically for each dimension. Extensive experiments on the public VIPL-HR and UBFC-rPPG datasets show that our proposed EST-TSANet outperforms the state-of-the-art (SOTA) methods.
Abstract The frequency and location distribution of tropical depressions (TDs) from 1979 to 2017 in the South China Sea (SCS) are statistically analyzed based on the best track data of tropical cyclones (TCs) from the Shanghai Typhoon Institute, China Meteorological Administration (CMA-STI). ECMWF interim reanalysis data (ERA-Interim) are used to investigate the reasons for the weakening of TDs in this study. The results show that there are 4.8 TDs formed in the SCS per year, and these TDs can be separated into 3.2 developing cases (DTDs) and 1.6 nondeveloping cases (NTDs) according to whether they intensify into tropical storms. Further objective classification by the multivariable-time empirical orthogonal function (MVT-EOF) method finds that the weakening cases in the positive-PC1 (the first principle component) mode occur in May–September, with the reason for weakening being a shortage of moisture. The decrease of westerly wind south of the NTDs reduces the water vapor transportation from the Indian Ocean. Binary TCs in the northwestern Pacific acquire water vapor from the eastern boundary of the SCS NTDs. Meanwhile, the weak high-level divergence and low-level convergence are not enough for the accumulation of local moisture and maintaining local convections inside the NTDs. The weakening cases in negative-PC1 mode occur in October–December with the reason for weakening being the invasion of cold air from the north. Strong cold air advection in the lower troposphere increases the vertical wind shear in front of the NTDs, and sharply reduce sensible and latent heat flux as well. Seasonal dependence exists in the causes of the SCS NTDs weakening.
Abstract In October 2022, an extreme cyclone developed in the South Pacific Ocean with a sea level pressure of 900 hPa, becoming the strongest extratropical cyclone in the satellite era. Using ERA5 reanalysis data, we investigated its development mechanisms and examined long‐term changes in the occurrence of extreme cyclones over the Southern Ocean. Our findings indicate that the cyclone formed within a low‐pressure anomaly over the South Pacific. Its explosive development was initiated by upper‐level dynamic forcing and driven by low‐level latent heat release, which had been preconditioned by surface heat flux. Extreme cyclones have increased significantly in the Amundsen‐Bellingshausen Seas (ABS) and the South Indian Ocean since 1980. The large‐scale environmental variables in the ABS also showed a consistent trend toward more favorable conditions for cyclone intensification. Understanding these extreme cyclone events will help to overcome the uncertainty in projections of climate change impacts and improve weather forecast skills.
For autonomous driving, one of the major challenges is to predict pedestrian crossing intention in ego-view. Pedestrian intention depends not only on their intrinsic goals but also on the stimulation of surrounding traffic elements. Considering the influence of other traffic elements on pedestrian intention, recent work introduced more traffic element information into the model to successfully improve performance. However, it is still difficult to effectively capture and fully exploit the potential dynamic spatio-temporal interactions among the target pedestrian and its surrounding traffic elements for accurate reasoning. In this work, inspired by neuroscience that human drivers tend to make continuous sensory-motor driving decisions by progressive visual stimulation, we propose a model termed Progressive Interaction Transformer (PIT) for pedestrian crossing intention prediction. Local pedestrian, global environment, and ego-vehicle motion are considered simultaneously in the proposed PIT. In particular, the temporal fusion block and self-attention mechanism are introduced to jointly and progressively model the dynamic spatio-temporal interactions among the three parties, allowing it to capture richer information and make prediction in a similar way to human drivers. Experimental results demonstrate that PIT achieves higher performance compared with other state-of-the-arts and preserves real-time inference.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)