Focus and track : pixel-wise spatio-temporal hurricane tracking

We proposes a pixel-wise extreme climate event tracking framework to track a target in the multiple moving objects scenario. We applied our model to tackle the challenging hurricane tracking problem. The proposed framework consists of two sub-models based on multi-layered ConvLSTM: a focus learning and a tracking model. Focus learning model learns location and appearance of target at first frame of video with auto-encoding fashion, and then, learned feature is fed into tracking model to follows the target in consecutive time frames. Extensive experiments show that the proposed tracking framework significantly outperforms against state-of-the-art tracking algorithms.

[1]  Andrew C. Lorenc,et al.  Analysis methods for numerical weather prediction , 1986 .

[2]  Harry C. Weber.  Hurricane track prediction using a statistical ensemble of numerical models , 2003 .

[3]  Fuqing Zhang,et al.  A Probabilistic Analysis of the Dynamics and Predictability of Tropical Cyclogenesis , 2008 .

[4]  R. Elsberry,et al.  Weighted Position and Motion Vector Consensus of Tropical Cyclone Track Prediction in the Western North Pacific , 2008 .

[5]  S. Majumdar,et al.  On the Ability of Global Ensemble Prediction Systems to Predict Tropical Cyclone Track Probabilities , 2010 .

[6]  Nathan Lay,et al.  Multimodel Approach Based on Evidence Theory for Forecasting Tropical Cyclone Tracks , 2010 .

[7]  Z. Pu,et al.  Tracking and Verification of East Atlantic Tropical Cyclone Genesis in the NCEP Global Ensemble: Case Studies during the NASA African Monsoon Multidisciplinary Analyses , 2010 .

[8]  T. T. Tien,et al.  Two-Dimensional Retrieval of Typhoon Tracks from an Ensemble of Multimodel Outputs , 2012 .

[9]  Hui Yu,et al.  Selective ensemble‐mean technique for tropical cyclone track forecast by using ensemble prediction systems , 2014 .

[10]  Prabhat,et al.  Application of Deep Convolutional Neural Networks for Detecting Extreme Weather in Climate Datasets , 2016, ArXiv.

[11]  T. T. Tien,et al.  Application of breeding ensemble to tropical cyclone track forecasts using the Regional Atmospheric Modeling System (RAMS) model , 2016 .

[12]  Fuqing Zhang,et al.  OBEST: An Observation-Based Ensemble Subsetting Technique for Tropical Cyclone Track Prediction , 2016 .

[13]  Prabhat,et al.  ExtremeWeather: A large-scale climate dataset for semi-supervised detection, localization, and understanding of extreme weather events , 2016, NIPS.

[14]  Wonjik Kim,et al.  Time Series Prediction of Tropical Storm Trajectory Using Self-Organizing Incremental Neural Networks and Error Evaluation , 2018, J. Adv. Comput. Intell. Intell. Informatics.

[15]  Sam Ganzfried,et al.  Predicting Hurricane Trajectories using a Recurrent Neural Network , 2018, AAAI.

[16]  Hyojin Kim,et al.  Deep-Hurricane-Tracker: Tracking and Forecasting Extreme Climate Events , 2019, 2019 IEEE Winter Conference on Applications of Computer Vision (WACV).

[17]  Sookyung Kim,et al.  Deep-dust: Predicting concentrations of fine dust in Seoul using LSTM , 2019, ArXiv.