Discriminative fine-grained network for vehicle re-identification using two-stage re-ranking

Wang, Qi; Min, Weidong; He, Daojing; Zou, Song; Huang, Tiemei; Liu, Ruikang

doi:10.1007/s11432-019-2811-8

Cited by 18 publications

(9 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the development of deep learning methods [17,18], and the crowd scenes become more crowded. Even the person in the image cannot be detected by object detection methods.…”

Section: Map-based Methods For Crowd Countingmentioning

confidence: 99%

MSR‐FAN: Multi‐scale residual feature‐aware network for crowd counting

Zhao

Min

Wei

et al. 2021

IET Image Processing

Self Cite

View full text Add to dashboard Cite

Crowd counting aims to count the number of people in crowded scenes, which is important to the security systems, traffic control and so on. The existing methods typically using local features cannot properly handle the perspective distortion and the varying scales in congested scene images, and henceforth perform wrong people counting. To alleviate this issue, this study proposes a multi-scale residual feature-aware network (MSR-FAN) that combines multi-scale features using multiple receptive field sizes and learns the featureaware information on each image. The MSR-FAN is trained end-to-end to generate highquality density map and evaluate the crowd number. The method consists of three parts. To handle the perspective changes problem, the first part, the direction-based featureenhanced network, is designed to encode the perspective information in four directions based on the initial image feature. The second part, the proposed multi-scale residual block module, gets the global information to handle the represent the regional feature better. This module explores features of different scales as well as reinforce the global feature. The third part, the feature-aware block, is designed to extract the feature hidden in the different channels. Experiment results based on benchmark datasets show that the proposed approach outperforms the existing state-of-the-art methods.

show abstract

“…With the development of deep learning methods [17,18], and the crowd scenes become more crowded. Even the person in the image cannot be detected by object detection methods.…”

Section: Map-based Methods For Crowd Countingmentioning

confidence: 99%

MSR‐FAN: Multi‐scale residual feature‐aware network for crowd counting

Zhao

Min

Wei

et al. 2021

IET Image Processing

Self Cite

View full text Add to dashboard Cite

show abstract

“…e metric learning method is mainly based on Markov metric learning (KISSME) and cross-view quadratic discriminant analysis (XQDA) [26]. For reidentification of human individuals, Zhong et al [27] proposed the k-reciprocal encoding algorithm, introduced the Jaccard distance, and merged it with the initial distance to improve the reidentification results.…”

Section: Reranking In Person and Vehicle Re-idmentioning

confidence: 99%

“…e reordering technology of vehicle reidentification has also attracted attention. Wang et al [26] used the k-reciprocal encoding algorithm to propose a discriminative fine-grained vehicle reidentification network based on a two-stage reordering framework. In the first stage, the k-reciprocal encoding feature is obtained from the fusion feature.…”

Section: Reranking In Person and Vehicle Re-idmentioning

confidence: 99%

“…Based on a deep learning network, GoogLeNet [32] was a neural network architecture that learns vehicle features and is pretrained on ImageNet. Different from conventional deep networks, two-branch convolutional networks were applied in PROVID [33], FACT + Plate-SNN + STR [34], Siamese-CNN [35], and DFN [26]. FACT + Plate-SNN + STR [34] was based on FACT, adding license plate information and spatiotemporal relationships to achieve distinguishing characteristics.…”

Section: Performance On the Veri-776 Datasetmentioning

confidence: 99%

“…FACT + Plate-SNN + STR [34] was based on FACT, adding license plate information and spatiotemporal relationships to achieve distinguishing characteristics. DFN [26] used fine-grained network to obtain global and local features for vehicle identification. STR + ST [36] and PAMA [25] separately process the key points of the vehicle.…”

Section: Performance On the Veri-776 Datasetmentioning

confidence: 99%

See 2 more Smart Citations

Vehicle Reidentification Based on MAPANet and k-Reciprocal Encoding

Tian

2021

Scientific Programming

View full text Add to dashboard Cite

Vehicle reidentification has important applications in intelligent monitoring systems. However, due to many factors, such as inaccurate vehicle image detection and viewing angle changes, distinguishing features cannot be effectively obtained when the vehicle is reidentified. To improve the recognition ability and robustness of vehicle reidentification, this study proposes a new multiattention part alignment network (MAPANet). The network uses different channels in the feature map to perceive different characteristics of the image clustering of the channels and achieves fine-grained attention to the vehicle. It can automatically locate the distinguishing subregions in the vehicle image and avoid the need for a large number of additional manual pretreatment steps. Moreover, an unsupervised reranking method based on multiple metrics is proposed. The k-reciprocal encoding algorithm can optimize the performance of the sorted list in the reordering problem, recalculate the interclass and intraclass distances of vehicle pictures, and improve sorting results. The experiments in this paper are carried out on the VeRi-776 and VehicleID datasets, and the mean average precision (mAP) results on the two datasets are 72.83% and 75.25%, respectively.

show abstract

Community detection by influential nodes based on random walk distance

Mokhtari,

Kherad

2024

Concurrency and Computation

View full text Add to dashboard Cite

SummaryCommunity detection is a key feature that can be used to extract useful information from networks. In most studies, the same static algorithms are used on real‐time snapshots of the dynamic network. Such an action increases the calculations and the time taken for the clustering operation. The idea of community detection is based on the fact that communities are formed around a node that is more popular and influential. Therefore, in the proposed algorithm, first, several snapshots are received from the network, then for the first snapshot, the influence of each node is calculated using the influence function based on the random walk distance between nodes. Then by selecting k nodes with higher influence, network communities are formed and other nodes belong to the community with the most common edges. In the next step, the next snapshots will be received and then the communities will be updated. Then K nodes with higher influence are selected and their community is created if needed. The proposed method has been compared with state‐of‐the‐art algorithms, and the results show that proposed method has been able to have a suitable performance in the uniformity of communities and also the speed of community formation.

show abstract

Discriminative fine-grained network for vehicle re-identification using two-stage re-ranking

Cited by 18 publications

References 48 publications

MSR‐FAN: Multi‐scale residual feature‐aware network for crowd counting

MSR‐FAN: Multi‐scale residual feature‐aware network for crowd counting

Vehicle Reidentification Based on MAPANet and k-Reciprocal Encoding

Community detection by influential nodes based on random walk distance

Contact Info

Product

Resources

About