Person reidentification using deep foreground appearance modeling

Watson, Gregory A.; Bhalerao, Abhir

doi:10.1117/1.jei.27.5.051215

Cited by 7 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Vehicle re-ID, a variation of person re-ID [9], has received increasing attention in recent years, as its viability continues to improve. Zapletal and Herout [10] were the first to collect a large-scale dataset for this purpose, conducting a vision-based study by utilizing color and oriented gradient histograms.…”

Section: Vehicle Re-id Methodsmentioning

confidence: 99%

Two-Stage Progressive Learning for Vehicle Re-Identification in Variable Illumination Conditions

Wu,

Jin,

2023

Electronics

View full text Add to dashboard Cite

Vehicle matching in variable illumination environments can be challenging due to the heavy dependence of vehicle appearance on lighting conditions. To address this issue, we propose a two-stage progressive learning (TSPL) framework. In the first stage, illumination-aware metric learning is enforced using a two-branch network via two illumination-specific feature spaces, used to explicitly model differences in lighting. In the second stage, discriminative feature learning is introduced to extract distinguishing features from a given vehicle. This process consists of a local feature extraction attention module, a local constraint, and a balanced sampling strategy. During the metric learning phase, the model expresses the union of local features, extracted from the attention module, with illumination-specific global features to form joint vehicle features. As part of the study, we construct a large-scale dataset, termed VERI-DAN (vehicle re-identification across day and night), to address the current lack of vehicle datasets exhibiting variable lighting conditions. This set is composed of 200,004 images from 16,654 vehicles, collected in various natural illumination environments. Validation experiments conducted with the VERI-DAN and Vehicle-1M datasets demonstrated that our proposed methodology effectively improved vehicle re-identification Rank-1 accuracy.

show abstract

Section: Vehicle Re-id Methodsmentioning

confidence: 99%

Two-Stage Progressive Learning for Vehicle Re-Identification in Variable Illumination Conditions

Wu,

Jin,

2023

Electronics

View full text Add to dashboard Cite

show abstract

“…Some of these techniques focus on extracting features that are both discriminative and invariant to various challenges. [5][6][7][8][9][10][11][12] However, it is challenging to design features that are discriminative enough to distinguish people reliably and at the same time invariant to factors such as motion blur, view angle, pose change, and other factors. To this end, the use of video is a desirable approach to improve the performance of reidentification.…”

Section: Introductionmentioning

confidence: 99%

CARF-Net: CNN attention and RNN fusion network for video-based person reidentification

et al. 2019

View full text Add to dashboard Cite

Video-based person reidentification is a challenging and important task in surveillance-based applications. Toward this, several shallow and deep networks have been proposed. However, the performance of existing shallow networks does not generalize well on large datasets. To improve the generalization ability, we propose a shallow end-to-end network which incorporates two stream convolutional neural networks, discriminative visual attention and recurrent neural network with triplet and softmax loss to learn the spatiotemporal fusion features. To effectively use both spatial and temporal information, we apply spatial, temporal, and spatiotemporal pooling. In addition, we contribute a large dataset of airborne videos for person reidentification, named DJI01. It includes various challenging conditions, such as occlusion, illuminationchanges, people with similar clothes, and the same people on different days. We perform elaborate qualitative and quantitative analyses to demonstrate the robust performance of the proposed model.

show abstract

“…3) Descriptor design. Data obtained by detection and tracking modules are used to segment the person's image for generating the descriptor, which can get constructed from data/cues as face [9]- [13]; visual appearance of the whole body [8], [14]- [21]; walking pattern [22], [23]; height and build [1]; and head, torso and limbs of a person [24], [25]; or a combination of these cues. 4) Descriptors matching.…”

Section: Introductionmentioning

confidence: 99%

Target Field of View Prediction Using Artificial Pheromones for People Reidentification

et al. 2019

View full text Add to dashboard Cite

People reidentification is a fundamental task in automated video surveillance based on computer vision. Reidentification happens when a person seen in a field of view is the same that has been observed in other fields of view. A person who has disappeared from one field of view can appear in any other within a camera network. Instead of looking for the person in all neighboring fields of view, for an intelligent video surveillance system, it is more practical to predict which of the neighboring camera views the person could appear. This prediction can become achieved by learning the paths the person usually follows in the camera network. The ant colony optimization technique has properties that can get exploited for this purpose; precisely, the accumulation and evaporation of artificial pheromones are used to learn the paths. After the learning process, the proposed method can make predictions every time that the person leaves a field of view. Such prediction is evaluated to obtain feedback and further tune the learning process. The path followed by the person becomes obtained by tracking their face image within and between fields of view using correlation filters as descriptors. The results obtained from an extensive experiment show that the field of view that the person selects to visit can be successfully predicted using artificial pheromones, and thus, reduce the resources that require reidentification.

show abstract

Person reidentification using deep foreground appearance modeling

Cited by 7 publications

References 19 publications

Two-Stage Progressive Learning for Vehicle Re-Identification in Variable Illumination Conditions

Two-Stage Progressive Learning for Vehicle Re-Identification in Variable Illumination Conditions

CARF-Net: CNN attention and RNN fusion network for video-based person reidentification

Target Field of View Prediction Using Artificial Pheromones for People Reidentification

Contact Info

Product

Resources

About