Kalman filter and Iterative-Hungarian Algorithm implementation for low complexity point tracking as part of fast multiple object tracking system

Sahbani, Bima; Adiprawita, Widyawardana

doi:10.1109/icsengt.2016.7849633

Cited by 50 publications

(17 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Various studies have been performed object matching based on the SURF features [42,43], Harris corner [44,45], and Hungarian method [46,47]. In Hungarian method, the objects are matched based on their optimal distance from each other.…”

Section: Simulation Resultsmentioning

confidence: 99%

Improved Object Matching in Multi-Objects Tracking Based On Zernike Moments and Combination of Multiple Similarity Metrics

Dadgar

Baleghi

Ezoji

2021

IJE

View full text Add to dashboard Cite

In video surveillance, multiple objects tracking (MOT) is a challenging task due to object matching problem in consecutive frames. The present paper aims to propose an improved object matching approach in MOT based on Zernike moments and combination of multiple similarity distance metrics. In this work, the object is primarily detected using background subtraction method while the Gaussian Mixture Model (GMM) is applied for object extraction in the next frames. Subsequently, the color histogram and the magnitude of Zernike moments of the objects are calculated. In the next step, the objects are matched in the current and the previous frames based on the Hausdorff distance between objects, Earth Mover's distance (EMD) between their color histograms, and Chi-square distance between their Zernike moments. Then, a voting mechanism is designed to find the best consensus object matching from the aforementioned metrics. Eventually, the location of each object is predicted by the Kalman filter to continue tracking in subsequent frames. The results show that the object tracking and matching performance is improved using the proposed method in the video sequences of the multi-camera pedestrian "EPFL" video dataset. Specifically, errors caused by the merging of targets are reduced in the proposed tracking process.

show abstract

Section: Simulation Resultsmentioning

confidence: 99%

Improved Object Matching in Multi-Objects Tracking Based On Zernike Moments and Combination of Multiple Similarity Metrics

Dadgar

Baleghi

Ezoji

2021

IJE

View full text Add to dashboard Cite

show abstract

“…Another challenge is the association of detection results from different video frames. Traditional approaches utilize the Hungarian algorithm [16] to check if the object in the current frame is the same as in the previous frame. However, the input to the Hungarian method is only the bounding box information, and thus may result in an unsolved matched due to the limited information.…”

Section: First Stage -Local Fusion and Associationmentioning

confidence: 99%

“…Stage -Track Fusion and Association. The Hungarian algorithm [16] is often used as the matching step for the trackers, which checks if the object in the current frame is the same as in the previous frame. For the set of a given detector matrix m from the CNN network and tracker matrix from the Kalman filter, it builds the association vector [ , ].…”

Section: Secondmentioning

confidence: 99%

Deep Associated Elastic Tracker for Intelligent Traffic Intersections

Liu

2020

Proceedings of the 2nd International Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Thi

View full text Add to dashboard Cite

Smart and connected traffic intersections are a key component of the smart city idea. These intersections will be equipped with dynamic street and traffic lights, traffic analysis, anomaly detection, and other "smart" features, contributing towards solving various energy, safety, and congestion problems. The objective of this paper is to lay out a process for utilizing deep learning technologies to develop a smart traffic flow for cities. With the increase in accurate deep learning video analytic models, we propose developing multiple object detection and tracking solutions for traffic video analytics. One key challenge is that the detection results from each video frame are not associated with each other, thus causing the traffic tracking to be lack continuity. To solve this problem, we propose a Deep Associated Elastic Tracker (DAE-Tracker) that performs multi-stage object association and tracking for various deeplearning based detectors. We tested our solution with multiple deep learning models under different datasets. We also deployed our system on the roof of our engineering building and contributed our own dataset. CCS CONCEPTS • Computer systems organization → Embedded systems; Sensor networks.

show abstract

“…t and c k t are k-th detection and candidate in frame t, respectively. To connect the candidate and detection within inter-frames, we match the candidates c k t and d k t in a bipartite graph with Hungarian algorithm [21]. The bi-…”

Section: Tracklet Generationmentioning

confidence: 99%

Trajectory Factory: Tracklet Cleaving and Re-Connection by Deep Siamese Bi-GRU for Multiple Object Tracking

Ma¹,

Yang²,

Yang³

et al. 2018

2018 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

Multi-Object Tracking (MOT) is a challenging task in the complex scene such as surveillance and autonomous driving. In this paper, we propose a novel tracklet processing method to cleave and re-connect tracklets on crowd or longterm occlusion by Siamese Bi-Gated Recurrent Unit (GRU). The tracklet generation utilizes object features extracted by CNN and RNN to create the high-confidence tracklet candidates in sparse scenario. Due to mis-tracking in the generation process, the tracklets from different objects are split into several sub-tracklets by a bidirectional GRU. After that, a Siamese GRU based tracklet re-connection method is applied to link the sub-tracklets which belong to the same object to form a whole trajectory. In addition, we extract the tracklet images from existing MOT datasets and propose a novel dataset to train our networks. The proposed dataset contains more than 95160 pedestrian images. It has 793 different persons in it. On average, there are 120 images for each person with positions and sizes. Experimental results demonstrate the advantages of our model over the state-of-the-art methods on MOT16.

show abstract

Kalman filter and Iterative-Hungarian Algorithm implementation for low complexity point tracking as part of fast multiple object tracking system

Cited by 50 publications

References 2 publications

Improved Object Matching in Multi-Objects Tracking Based On Zernike Moments and Combination of Multiple Similarity Metrics

Improved Object Matching in Multi-Objects Tracking Based On Zernike Moments and Combination of Multiple Similarity Metrics

Deep Associated Elastic Tracker for Intelligent Traffic Intersections

Trajectory Factory: Tracklet Cleaving and Re-Connection by Deep Siamese Bi-GRU for Multiple Object Tracking

Contact Info

Product

Resources

About