Cross-modal distillation for RGB-depth person re-identification

Hafner, Frank M.; Bhuyian, Amran; Kooij, Julian F. P.; Granger, Éric

doi:10.1016/j.cviu.2021.103352

Cited by 23 publications

(10 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Single-modal person Re-ID matches probe samples with gallery samples, and all samples are taken from the same modality (i.e., RGB–RGB or IR–IR matching) [ 65 ]. Unlike single-modal Re-ID, cross-modal Re-ID aims to match the probe sample taken from one modality against a gallery set from another modality, such as RGB–IR, RGB–depth, or sketch–RGB images [ 13 , 52 , 66 ], as shown in Figure 10 .…”

Section: Cross-modal Person Re-identificationmentioning

confidence: 99%

“…RGB–depth cross-modal Re-ID matches a person of interest across RGB and depth modalities. Recently, a few works [ 49 , 50 , 51 , 52 , 53 ] investigated how to mitigate the domain gap between RGB and depth modalities. As the nature of the data for both modalities is heterogeneous, Re-ID researchers have proposed their own methods to tackle the challenges of the process, such as data discrepancy, and have utilized conventional [ 49 , 50 ] and deep-learning methods [ 51 , 52 , 53 ] to identify individuals across the two modalities.…”

Section: Cross-modal Person Re-identificationmentioning

confidence: 99%

“…Recently, Hafner et al [ 51 , 52 ] proposed a deep transfer learning technique where a cross-modal distillation network was employed to transfer knowledge from RGB modalities to depth modalities or vice versa to solve cross-modal Re-ID between RGB and depth. The authors extended the work of [ 51 ] by changing embedding size as a hyperparameter for the optimization of Re-ID to achieve a superior accuracy [ 52 ]. They evaluated their method using two public datasets, BIWI RGBD-ID and RobotPKU, and obtained a 42.8 ± 3.9 % rank-1 recognition for BIWI with RGB as the gallery and depth as the query.…”

Section: Cross-modal Person Re-identificationmentioning

confidence: 99%

“…Similarly, rank-1 accuracies of 16.32% and 28.98% with the BIWI RGBD-ID dataset were achieved, and 27.95% and 32.09% with the IAS Lab dataset. In [ 52 ], Hafner et al proposed the cross-modal person Re-ID method discussed in Section 3.2 . The authors also investigated depth–depth single-modal Re-ID and evaluated its performance with BIWI and RobotPKU datasets.…”

Section: Single-modal Person Re-idmentioning

confidence: 99%

“…In the Re-ID literature, publicly available RGB–D and IR sensor-based datasets [ 4 , 14 , 33 , 35 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 ] have been recorded for specific purposes because some Re-ID approaches use different domains combinedly to improve the performance of Re-ID [ 33 , 34 , 35 , 36 , 37 , 38 , 39 , 47 , 48 ] while other approaches use cross-domain methods [ 49 , 50 , 51 , 52 , 53 ] in the Re-ID framework to continue surveillance between daytime visible images and nighttime infrared images. Still, some Re-ID methods consider only IR–IR [ 17 , 54 ] matching using the publicly available datasets.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Person Re-Identification with RGB–D and RGB–IR Sensors: A Comprehensive Survey

Uddin

Bhuiyan

Bappee

et al. 2023

Sensors

View full text Add to dashboard Cite

Learning about appearance embedding is of great importance for a variety of different computer-vision applications, which has prompted a surge in person re-identification (Re-ID) papers. The aim of these papers has been to identify an individual over a set of non-overlapping cameras. Despite recent advances in RGB–RGB Re-ID approaches with deep-learning architectures, the approach fails to consistently work well when there are low resolutions in dark conditions. The introduction of different sensors (i.e., RGB–D and infrared (IR)) enables the capture of appearances even in dark conditions. Recently, a lot of research has been dedicated to addressing the issue of finding appearance embedding in dark conditions using different advanced camera sensors. In this paper, we give a comprehensive overview of existing Re-ID approaches that utilize the additional information from different sensor-based methods to address the constraints faced by RGB camera-based person Re-ID systems. Although there are a number of survey papers that consider either the RGB–RGB or Visible-IR scenarios, there are none that consider both RGB–D and RGB–IR. In this paper, we present a detailed taxonomy of the existing approaches along with the existing RGB–D and RGB–IR person Re-ID datasets. Then, we summarize the performance of state-of-the-art methods on several representative RGB–D and RGB–IR datasets. Finally, future directions and current issues are considered for improving the different sensor-based person Re-ID systems.

show abstract