End-to-End Chained Pedestrian Multi-Object Tracking Based on Multi-Feature Fusion

Zhou, Haiyun; Xiang, Xuezhi; Wang, Xinyao; Ren, Wenkai

doi:10.1109/icsess52187.2021.9522289

Cited by 3 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Four samples were randomly selected from the KITTI test set for testing. The smoothness of the results obtained by each algorithm was calculated using the above formula (26), and the results are shown in Table 4.…”

Section: Test Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Design of Reliable Mining Algorithm for Massive Moving Image Data Trajectory

Zeng,

Run,

Djenouri

2024

IEEE Access

View full text Add to dashboard Cite

To acquire precise, dependable, and credible trajectory information from extensive motion image datasets, this study introduces a robust mining algorithm grounded in trajectory extraction for voluminous motion data. The algorithm leverages an enhanced single-stage object detection model (TFF-SSD) and employs a 3D sparse convolutional neural network for extracting point cloud features from the extensive motion data. Simultaneously, spatial semantic features are derived by integrating spatial and semantic characteristics of the vast motion data. Subsequently, a mutual attention fusion method is applied to amalgamate point cloud features and spatial semantic features into the Faster R-CNN model. This facilitates the swift identification of moving target regions, leading to the extraction of classification and coordinate information from the abundant data. Employing reference points from diverse types of extensive motion data, the algorithm selects associated feature regions for the motion data target, thereby obtaining reference point information indicative of continuous motion. Subsequent fitting of this reference point information enables the reliable mining of trajectories within the extensive motion data. Experimental outcomes demonstrate the algorithm's ability to accurately detect all crowd movements in a running race. The extracted motion data features are diverse, encompassing reference points that facilitate trajectory mining with elevated quality, reliability, and genuine data content.

show abstract

Section: Test Resultsmentioning

confidence: 99%

“…In the formula, Lcls and Lreg represent classification loss function and regression loss function respectively, and Lcls is represented by FocalLoss function [26][27] shown in Equation 11. For the 3D target detection of massive motion data, it is composed of 7 parameters.…”

Section: ) Loss Functionmentioning

confidence: 99%

Design of Reliable Mining Algorithm for Massive Moving Image Data Trajectory

Zeng,

Run,

Djenouri

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Non-rigid 3D image reconstruction seeks to recover a 3D model of non-rigid objects from 2D images captured from multiple viewpoints, employing image processing and computer vision techniques [4]. This technology has wideranging applications in movie production, game development, and industrial design [5,6]. However, the complexity and diversity of non-rigid motion render non-rigid 3D image reconstruction a challenging task, with ensuring the reliability of shape bases computation standing out as a particular concern.…”

Section: Introductionmentioning

confidence: 99%

A Non-Rigid Three-Dimensional Image Reconstruction Algorithm Based on Deformable Shape Reliability

Chen,

Moqurrab

2024

IEEE Access

View full text Add to dashboard Cite

Most reconstruction algorithms for non-rigid three-dimensional (3D) images assume that non-rigidity can be represented as a linear combination of a fixed number of rigid bases. However, this assumption struggles to establish reliable shape functions and initial values for nonlinear and non-rigid motions, decreasing reconstruction reliability. This paper introduces an enhanced-reliability reconstruction algorithm for non-rigid 3D images. Our algorithm models the dynamic non-rigid shape basis as a low-rank matrix composed of image points and depth factors, improving the restoration of non-rigid shape base changes and providing accurate parameters for constructing objective functions. By leveraging manifold alignment and physical continuity constraints, our method optimizes the function structures. Assuming minimal reconstruction error and shape change, we solve for the motion structure parameters and select the key initial shape basis value by minimizing the objective function with the L-M nonlinear optimization method. Our experimental results on 3D image sequence reconstructions demonstrate significant error reduction, underscoring our model's credibility, robust reliability, and minimal re-projection error.

show abstract