Zhiqiang Jiang

Zhiqiang Jiang

Auto-generated profile, based on publicly available information

Zhiqiang Jiang

State Key Laboratory of Robotics and System, Harbin Institute of Technology, Harbin, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 October 2023
Verified
Conference Paper
Open access
XLXiaoyu Li
Xiaoyu Li
TXTao Xie
Tao Xie
DLDedong Liu
Dedong Liu
JGJinghan Gao
Jinghan Gao
KDKun Dai
Kun Dai
WKWang Ke
Wang Ke
3D Multi-object tracking (MOT) empowers mobile robots to accomplish well-informed motion planning and navigation tasks by providing motion trajectories of surrounding objects. However, existing 3D MOT methods typically employ a single similarity metric and physical model to perform data association and state estimation for all objects. With large-scale modern datasets and real scenes, there are a variety of object categories that commonly exhibit distinctive geometric properties and motion patterns. In this way, such distinctions would enable various object categories to behave differently under the same standard, resulting in erroneous matches between trajectories and detections, and jeopardizing the reliability of downstream tasks (navigation, etc.). Towards this end, we propose Poly-MOT, an efficient 3D MOT method based on the Tracking-By-Detection framework that enables the tracker to choose the most appropriate tracking criteria for each object category. Specifically, Poly-MOT leverages different motion models for various object categories to characterize distinct types of motion accurately. We also introduce the constraint of the rigid structure of objects into a specific motion model to accurately describe the highly nonlinear motion of the object. Additionally, we introduce a two-stage data association strategy to ensure that objects can find the optimal similarity metric from three custom metrics for their categories and reduce missing matches. On the NuScenes dataset, our proposed method achieves state-of-the-art performance with 75.4% AMOTA. The code is available at https://github.com/lixiaoyu20001P0Iy-MOT.

21 December 2023 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
TXTao Xie
Tao Xie
KDKun Dai
Kun Dai
ZJZhiqiang Jiang
Zhiqiang Jiang
RLRuifeng Li
Ruifeng Li
SMShouren Mao
Shouren Mao
LZLijun Zhao
Lijun Zhao
In this work, we seek to learn multiple mainstream vision tasks concurrently using a unified network, which is storage-efficient as numerous networks with task-shared parameters can be implanted into a single consolidated network. Our framework, vision transformer (ViT)-MVT, built on a plain and nonhierarchical ViT, incorporates numerous visual tasks into a modest supernet and optimizes them jointly across various dataset domains. For the design of ViT-MVT, we augment the ViT with a multihead self-attention (MHSE) to offer complementary cues in the channel and spatial dimension, as well as a local perception unit (LPU) and locality feed-forward network (locality FFN) for information exchange in the local region, thus endowing ViT-MVT with the ability to effectively optimize multiple tasks. Besides, we construct a search space comprising potential architectures with a broad spectrum of model sizes to offer various optimum candidates for diverse tasks. After that, we design a layer-adaptive sharing technique that automatically determines whether each layer of the transformer block is shared or not for all tasks, enabling ViT-MVT to obtain task-shared parameters for a reduction of storage and task-specific parameters to learn task-related features such that boosting performance. Finally, we introduce a joint-task evolutionary search algorithm to discover an optimal backbone for all tasks under total model size constraint, which challenges the conventional wisdom that visual tasks are typically supplied with backbone networks developed for image classification. Extensive experiments reveal that ViT-MVT delivers exceptional performances for multiple visual tasks over state-of-the-art methods while necessitating considerably fewer total storage costs. We further demonstrate that once ViT-MVT has been trained, ViT-MVT is capable of incremental learning when generalized to new tasks while retaining identical performances for trained tasks. The code is available at https://github.com/XT-1997/vitmvt.

27 February 2024 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
KDKun Dai
Kun Dai
WKWang Ke
Wang Ke
TXTao Xie
Tao Xie
TSTao Sun
Tao Sun
JZJinhang Zhang
Jinhang Zhang
MOMohamed Omar
Mohamed Omar
Local feature matching, which aims to establish the matches between image pairs, is a pivotal component of multiple visual applications. While current transformer-based works exhibit remarkable performance, they mechanically alternate self- and cross-attention in a predetermined order without considering their prioritization, culminating in inadequate enhancement of visual descriptors. Moreover, when calculating attention matrices to integrate global context, current methods only explicitly model the correlation among the feature channels without taking their importance into account, leaving insufficient message propagation. In this work, we develop a dynamic sparse attention perception (DSAP) matcher to tackle the aforementioned issues. To resolve the first issue, DSAP presents a dynamic perception strategy that enables the network to dynamically implement feature enhancement via modifying both forward and backward propagation. During forward propagation, DPS assigns a learnable perception score to each transformer layer and employs an exponential moving average algorithm to calculate the current score. After that, DPS utilizes an indicator function to binarize the score, allowing DSAP to adaptively determine the appropriate utilization of self- or cross-attention at current iteration. During backward propagation, DPS employs a gradient estimator that adjusts the gradient of perception scores, thus rendering them differentiable. To tackle the second issue, DSAP introduces a weighted sparse transformer (WSFormer) that recalibrates attention matrices by concurrently considering both channel importance and channel correlation. WSFormer predicts attention vectors to weight attention matrices while constructing multiple sparse attention matrices to integrate various global messages, thus highlighting informative channels and inhibiting redundant message propagation. Extensive experiments in public datasets and real environment demonstrate that DSAP achieves exceptional performances across various downstream tasks, including relative pose estimation and visual localization. The code is available at https://github.com/mooncake199809/DSAP.

26 August 2023 in Expert Systems with Applications
Verified
Journal Article
Closed access
KDKun Dai
Kun Dai
TXTao Xie
Tao Xie
WKWang Ke
Wang Ke
ZJZhiqiang Jiang
Zhiqiang Jiang
RLRuifeng Li
Ruifeng Li
LZLijun Zhao
Lijun Zhao

17 August 2023 in IEEE Transactions on Multimedia
Verified
Journal Article
Closed access
LWLi Wang
Li Wang
TXTao Xie
Tao Xie
XZXinyu Zhang
Xinyu Zhang
ZJZhiqiang Jiang
Zhiqiang Jiang
LYLinqi Yang
Linqi Yang
HLHuaping Liu
Huaping Liu
Pure point-based neural networks have recently shown tremendous promise for point cloud tasks, including 3D object classification, 3D object part segmentation, 3D semantic segmentation, and 3D object detection. Nevertheless, it is a laborious process to construct a network for each task due to the artificial parameters and hyperparameters involved, e.g., the depths and widths of the network and the number of sampled points at each stage. In this work, we propose Auto-Points, a novel one-shot search framework that automatically seeks the optimal architecture configuration for point cloud tasks. Technically, we introduce a set abstraction mixer (SAM) layer that is capable of scaling up flexibly along the depth and width of the network. Each SAM layer consists of numerous child candidates, which simplifies architecture search and enables us to discover the optimum design for each point cloud task pursuant to resource constraint from an enormous search space. To fully optimize the child candidates, we develop a weight-entwinement neural architecture search (NAS) technique that entwines the weights of different candidates in the same layer during supernet training such that all candidates can be extremely optimized. Benefiting from the proposed techniques, the trained supernet allows the searched subnets to be exceptionally well-optimized without further retraining or finetuning. In particular, the searched models deliver superior performances on multiple extensively employed benchmarks, 93.9% overall accuracy (OA) on ModelNet40, 89.1% OA on ScanObjectNN, 87.1% instance average IoU on ShapeNetPart, 69.1% mIoU on S3DIS, 70.4% mAP@0.25 on ScanNet V2, and 64.4% mAP@0.25 on SUN RGB-D.

14 September 2023 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
TXTao Xie
Tao Xie
ZJZhiqiang Jiang
Zhiqiang Jiang
SLShuozhan Li
Shuozhan Li
YZYukun Zhang
Yukun Zhang
KDKun Dai
Kun Dai
LZLijun Zhao
Lijun Zhao
Recently, scene coordinate regression (SCoRe) approaches for visual localization have been extensively investigated. However, current SCoRe methods are scene-specific and necessitate retraining when generalizing to new scenarios, leaving a consistent rise in model capacity as the amount of scenes increases. To this end, we develop UFVL-Net, a unifying framework that integrates localization tasks of multiple indoor scenarios into a manageable network and optimizes these tasks collectively under diversified scene domains, where the localization of each scenario domain is considered as a separate task. UFVL-Net is storage-efficient since multiple models with shared parameters can be consolidated into a single one. Specifically, we introduce two parameter sharing policies, i.e., channel-wise sharing policy and kernel-wise sharing policy, which offer fine-grained parameter sharing within each layer of the backbone for efficient storage while providing task-specific parameters to tackle the inherent hurdles associated with multi-domain learning for visual localization, i.e., gradient conflict due to a skewed competition among tasks for the shared parameters. The key insight lies in that leveraging task-sharing parameters can learn a generic feature representation across scenes while utilizing task-specific parameters can learn task-related features for alleviating gradient conflict. Moreover, we develop a sign-based gradient normalization technique applied to task-sharing parameters to promote the training of UFVL-Net by further mitigating gradient conflict, thus emphasizing the utilization of task-sharing parameters and ensuring that each task is thoroughly optimized. We undertake extensive experiments across numerous datasets and complex real-world scenarios, showing that UFVL-Net families significantly outperform the cutting-edge methods with much less storage space. We demonstrate UFVL-Net can be generalized to new scenarios using a few task-specific parameters, further highlighting the superiority of UFVL-Net. The code is available at here.