The Experts below are selected from a list of 324 Experts worldwide ranked by ideXlab platform
Shu Tang - One of the best experts on this subject based on the ideXlab platform.
-
real time tracking of vehicles with siamese network and Backward Prediction
International Conference on Multimedia and Expo, 2020Co-Authors: Lei Luo, Shu TangAbstract:Tracking of vehicles is a key technique for Intelligent transportation system, which commonly follows tracking-by-detection strategy. Due to high appearance similarity among vehicles and heavy occlusion caused by busy traffic flow, a major challenge in such a tracking system is the limited performance of the underlying detector which may produce noisy detections. Consequently, Siamese network and Backward Prediction-based vehicle tracking approach is proposed. Siamese network based forward position Prediction is designed to alleviate the interference of noisy detections, while Backward Prediction verification is performed to reduce the false positives arising with forward Prediction. The final tracklets are obtained through weighted merging based on the detection confidence and forward Prediction confidence. The experiment results demonstrate that the proposed method outperforms the state-of-the-art on the UA-DETRAC vehicle tracking dataset, as well as maintains real-time processing at an average tracking speed of 20.1fps, which can be used for real-time applications.
-
ICME - Real-Time Tracking of Vehicles with Siamese Network and Backward Prediction
2020 IEEE International Conference on Multimedia and Expo (ICME), 2020Co-Authors: Lei Luo, Shu TangAbstract:Tracking of vehicles is a key technique for Intelligent transportation system, which commonly follows tracking-by-detection strategy. Due to high appearance similarity among vehicles and heavy occlusion caused by busy traffic flow, a major challenge in such a tracking system is the limited performance of the underlying detector which may produce noisy detections. Consequently, Siamese network and Backward Prediction-based vehicle tracking approach is proposed. Siamese network based forward position Prediction is designed to alleviate the interference of noisy detections, while Backward Prediction verification is performed to reduce the false positives arising with forward Prediction. The final tracklets are obtained through weighted merging based on the detection confidence and forward Prediction confidence. The experiment results demonstrate that the proposed method outperforms the state-of-the-art on the UA-DETRAC vehicle tracking dataset, as well as maintains real-time processing at an average tracking speed of 20.1fps, which can be used for real-time applications.
V Cappellini - One of the best experts on this subject based on the ideXlab platform.
-
adaptively weighted vector median filters for motion fields smoothing
International Conference on Acoustics Speech and Signal Processing, 1996Co-Authors: Luciano Alparone, Mauro Barni, Franco Bartolini, V CappelliniAbstract:In the field of video coding the issues of Backward Prediction and standards conversion have focused an increasing attention towards techniques for an effective estimation of the true interframe motion. The problem of restoration of motion vector-fields computed by means of a standard block matching algorithm is addressed. The restoration must be carried out carefully by exploiting both the spatial correlation of the vector-field, and the significance of the obtained vectors as measures of the reliability of the previous estimation step. A novel approach matching both the above requirements is presented. Based on the theory of vector-median filters an adaptive scheme is developed and results are discussed.
-
ICASSP - Adaptively weighted vector-median filters for motion-fields smoothing
1996 IEEE International Conference on Acoustics Speech and Signal Processing Conference Proceedings, 1Co-Authors: Luciano Alparone, Mauro Barni, Franco Bartolini, V CappelliniAbstract:In the field of video coding the issues of Backward Prediction and standards conversion have focused an increasing attention towards techniques for an effective estimation of the true interframe motion. The problem of restoration of motion vector-fields computed by means of a standard block matching algorithm is addressed. The restoration must be carried out carefully by exploiting both the spatial correlation of the vector-field, and the significance of the obtained vectors as measures of the reliability of the previous estimation step. A novel approach matching both the above requirements is presented. Based on the theory of vector-median filters an adaptive scheme is developed and results are discussed.
Lei Luo - One of the best experts on this subject based on the ideXlab platform.
-
real time tracking of vehicles with siamese network and Backward Prediction
International Conference on Multimedia and Expo, 2020Co-Authors: Lei Luo, Shu TangAbstract:Tracking of vehicles is a key technique for Intelligent transportation system, which commonly follows tracking-by-detection strategy. Due to high appearance similarity among vehicles and heavy occlusion caused by busy traffic flow, a major challenge in such a tracking system is the limited performance of the underlying detector which may produce noisy detections. Consequently, Siamese network and Backward Prediction-based vehicle tracking approach is proposed. Siamese network based forward position Prediction is designed to alleviate the interference of noisy detections, while Backward Prediction verification is performed to reduce the false positives arising with forward Prediction. The final tracklets are obtained through weighted merging based on the detection confidence and forward Prediction confidence. The experiment results demonstrate that the proposed method outperforms the state-of-the-art on the UA-DETRAC vehicle tracking dataset, as well as maintains real-time processing at an average tracking speed of 20.1fps, which can be used for real-time applications.
-
ICME - Real-Time Tracking of Vehicles with Siamese Network and Backward Prediction
2020 IEEE International Conference on Multimedia and Expo (ICME), 2020Co-Authors: Lei Luo, Shu TangAbstract:Tracking of vehicles is a key technique for Intelligent transportation system, which commonly follows tracking-by-detection strategy. Due to high appearance similarity among vehicles and heavy occlusion caused by busy traffic flow, a major challenge in such a tracking system is the limited performance of the underlying detector which may produce noisy detections. Consequently, Siamese network and Backward Prediction-based vehicle tracking approach is proposed. Siamese network based forward position Prediction is designed to alleviate the interference of noisy detections, while Backward Prediction verification is performed to reduce the false positives arising with forward Prediction. The final tracklets are obtained through weighted merging based on the detection confidence and forward Prediction confidence. The experiment results demonstrate that the proposed method outperforms the state-of-the-art on the UA-DETRAC vehicle tracking dataset, as well as maintains real-time processing at an average tracking speed of 20.1fps, which can be used for real-time applications.
A Benaskeur - One of the best experts on this subject based on the ideXlab platform.
-
CDC - Forward Prediction-based approach to target-tracking with Out-of-Sequence Measurements
2008 47th IEEE Conference on Decision and Control, 2008Co-Authors: F Rheaume, A BenaskeurAbstract:In target-tracking applications, there may be situations where measurements from a given target arrive out of sequence at the processing center. This problem is commonly referred to as out-of-sequence measurements (OOSMs). So far, most of the existing solutions to this problem are based on retrodiction, where Backward Prediction of the current state is used to incorporate the OOSMs at the appropriate time. This paper suggests a new method for tackling the OOSMs problem without Backward Prediction. Based on a forward Prediction and de-correlation approach, the method has proved to be as performing as the best retrodiction-based methods, while requiring less data storage in most cases.
-
Out-of-Sequence Measurements Filtering Using Forward Prediction
2007Co-Authors: F Rheaume, A BenaskeurAbstract:Abstract : In target tracking applications, there may be situations where measurements from a given target arrive out of sequence at the processing center. This problem is commonly referred to as Out-of-Sequence Measurements (OOSMs). So far, most of the existing solutions to this problem are based on retrodiction, where Backward Prediction of the current estimated state is used to incorporate the OOSMs at appropriate time instants. This paper suggests a new method for tackling the OOSMs problem without Backward Prediction. Based on a forward Prediction and de-correlation approach, the method has proved to compare favorably to the best retrodiction-based methods, while requiring less data storage in most cases.
Luciano Alparone - One of the best experts on this subject based on the ideXlab platform.
-
adaptively weighted vector median filters for motion fields smoothing
International Conference on Acoustics Speech and Signal Processing, 1996Co-Authors: Luciano Alparone, Mauro Barni, Franco Bartolini, V CappelliniAbstract:In the field of video coding the issues of Backward Prediction and standards conversion have focused an increasing attention towards techniques for an effective estimation of the true interframe motion. The problem of restoration of motion vector-fields computed by means of a standard block matching algorithm is addressed. The restoration must be carried out carefully by exploiting both the spatial correlation of the vector-field, and the significance of the obtained vectors as measures of the reliability of the previous estimation step. A novel approach matching both the above requirements is presented. Based on the theory of vector-median filters an adaptive scheme is developed and results are discussed.
-
ICASSP - Adaptively weighted vector-median filters for motion-fields smoothing
1996 IEEE International Conference on Acoustics Speech and Signal Processing Conference Proceedings, 1Co-Authors: Luciano Alparone, Mauro Barni, Franco Bartolini, V CappelliniAbstract:In the field of video coding the issues of Backward Prediction and standards conversion have focused an increasing attention towards techniques for an effective estimation of the true interframe motion. The problem of restoration of motion vector-fields computed by means of a standard block matching algorithm is addressed. The restoration must be carried out carefully by exploiting both the spatial correlation of the vector-field, and the significance of the obtained vectors as measures of the reliability of the previous estimation step. A novel approach matching both the above requirements is presented. Based on the theory of vector-median filters an adaptive scheme is developed and results are discussed.