Appearance Feature Based Human Correspondence under Non-overlapping Views

Chae, Hyun-Uk; Jo, Kang-Hyun

doi:10.1007/978-3-642-04070-2_69

Cited by 5 publications

(11 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 4 lists the comparative matching parameters of this proposed method with Hyun-Uk's proposal in [5]. The figures show the increase of accuracy rate and the decrease of omit matching rate.…”

Section: Experimental Results In Simple Scenariomentioning

confidence: 99%

“…To verify the real-time of the appearance feature extraction in this paper, all the 898 frames captured from two views are preprocessed into 720×540 bmp images and their appearance features are extracted through the proposals of Hyun-Uk's [5] and Madden's [7] as well as this proposed method. The triple comparison diagram of computation time is illustrated in Fig.4.…”

Section: Experimental Results In Simple Scenariomentioning

confidence: 99%

“…The proposed method is further compared with other human tracking methods like Hyun-Uk's [5] Madden's [7] and Trevor's [11]. Table 5 lists the matching parameters of each method.…”

Section: Experimental Results In Complex Scenariosmentioning

confidence: 99%

“…As for feature extraction, previous literatures have concentrated on appearance features to track human objects across disjoint cameras [2][3][4][5][6] . Gianfranco et al [4] reviewed the development of a set of models that capture the overall appearance of an individual.…”

Section: Related Workmentioning

confidence: 99%

“…Some of the models provide a holistic description of a person, and some others require an intermediate step where specific body parts need to be identified. Hyun-Uk et al [5] segmented each person by a criterion with appearance and estimated the segmented regions as Gaussian mixture model (GMM) for correspondence. Both experimental results demonstrated that appearance feature has superior performance in identifying individuals, but illumination changes pose a major problem to the invariance of appearance features.…”

Section: Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Robust Human Tracking Method Based on Apperance and Geometrical Features in Non-Overlapping Views

Liang¹,

Han²,

Zhu³

et al. 2015

Computer Science &Amp; Information Technology ( CS &Amp; IT )

View full text Add to dashboard Cite

show abstract

Section: Experimental Results In Simple Scenariomentioning

confidence: 99%

Section: Experimental Results In Simple Scenariomentioning

confidence: 99%

“…The proposed method is further compared with other human tracking methods like Hyun-Uk's [5] Madden's [7] and Trevor's [11]. Table 5 lists the matching parameters of each method.…”

Section: Experimental Results In Complex Scenariosmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Robust Human Tracking Method Based on Apperance and Geometrical Features in Non-Overlapping Views

Liang¹,

Han²,

Zhu³

et al. 2015

Computer Science &Amp; Information Technology ( CS &Amp; IT )

View full text Add to dashboard Cite

show abstract

An Effective Adversarial Attack on Person Re-Identification in Video Surveillance via Dispersion Reduction

Zheng

Velipasalar

2020

IEEE Access

View full text Add to dashboard Cite

Person re-identification across a network of cameras, with disjoint views, has been studied extensively due to its importance in wide-area video surveillance. This is a challenging task due to several reasons including changes in illumination and target appearance, and variations in camera viewpoint and camera intrinsic parameters. The approaches developed to re-identify a person across different camera views need to address these challenges. More recently, neural network-based methods have been proposed to solve the person re-identification problem across different camera views, achieving state-of-the-art performance. In this paper, we present an effective and generalizable attack model that generates adversarial images of people, and results in very significant drop in the performance of the existing state-of-the-art person reidentification models. The results demonstrate the extreme vulnerability of the existing models to adversarial examples, and draw attention to the potential security risks that might arise due to this in video surveillance. Our proposed attack is developed by decreasing the dispersion of the internal feature map of a neural network to degrade the performance of several different state-of-the-art person re-identification models. We also compare our proposed attack with other state-of-the-art attack models on different person reidentification approaches, and by using four different commonly used benchmark datasets. The experimental results show that our proposed attack outperforms the state-of-art attack models on the best performing person re-identification approaches by a large margin, and results in the most drop in the mean average precision values.

show abstract

Analysis of Brightness Transfer Function for Matching Targets across Networked Cameras

Kumar

Doğançay

2011

2011 International Conference on Digital Image Computing: Techniques and Applications

View full text Add to dashboard Cite

In this paper we present an analysis of the intercamera matching problem based on colour and intensity of targets. Brightness Transfer Function (BTF) have been quite effective in matching targets across cameras with nonoverlapping fields of view. The BTF effectively compensates for the illumination changes in the environment and radiometric distortions due to difference in camera sensors. We used the Cumulative Brightness Transfer Function (CBTF) to model the relationship of separate radiometric responses of the cameras and difference in scene illumination. No shape assumption on the CBTF is made. A non-parametric and non-linear CBTF is computed from training data, where the identity of the targets across cameras are known. We studied the effect of the use of different shortest path algorithms in the computation of the BTF. Furthermore we analysed the effect of quantisation on the CBTF and its further effect on matching of the targets. We used the Bhattacarya distance to compute the match measure for reference and candidate transformed histograms. The effect of quantization is studied by quantitative evaluation in terms of temporal analysis, mean square error, and finally as Receiver Operating Characteristic (ROC) plots for matching targets across cameras in a network of cameras.

show abstract

Appearance Feature Based Human Correspondence under Non-overlapping Views

Cited by 5 publications

References 8 publications

Robust Human Tracking Method Based on Apperance and Geometrical Features in Non-Overlapping Views

Robust Human Tracking Method Based on Apperance and Geometrical Features in Non-Overlapping Views

An Effective Adversarial Attack on Person Re-Identification in Video Surveillance via Dispersion Reduction

Analysis of Brightness Transfer Function for Matching Targets across Networked Cameras

Contact Info

Product

Resources

About