Fusion in dissimilarity space for RGB-D person re-identification

Uddin, Kamal; Lam, Antony; Fukuda, Hisato; Kobayashi, Yoshinori; Kuno, Yoshinori

doi:10.1016/j.array.2021.100089

Cited by 10 publications

(11 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some works [ 37 , 38 ] combined RGB appearance features with illumination-invariant anthropometric features from processed depth images. Apart from the above works, some Re-ID methods use raw-depth and RGB images combinedly [ 40 , 42 , 47 ]. Moreover, in some approaches [ 48 , 57 ], depth frames are used as body segmentation masks, which play an important role in removing cluttered backgrounds from scenes.…”

Section: Multi-modal Person Re-identificationmentioning

confidence: 99%

“…In [ 42 ], the authors proposed a method to fuse RGB and RGB–D image-based ranking scores in a dissimilarity space for person re-identification. They trained two models using either RGB or RGB–D images and then computed dissimilarity scores between the probe and gallery sets using feature embeddings extracted from each model.…”

Section: Multi-modal Person Re-identificationmentioning

confidence: 99%

“…Finally, both dissimilarity scores were fused in the dissimilarity space to obtain the final recognition accuracy between the probe and gallery sets. Although the computational cost is high, this approach obtained superior accuracies on two publicly available datasets—RGBD-ID [ 58 ] and RobotPKU RGBD-ID [ 33 ]—and their own dataset, SUCVL RGBD-ID [ 42 ]. They achieved rank-1 accuracies of 93.33%, 82.05%, and 87.65% for RobotPKU, RGBD-ID, and SUCVL, respectively.…”

Section: Multi-modal Person Re-identificationmentioning

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 3 more Smart Citations

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

Uddin

Bhuiyan

Bappee

et al. 2023

Sensors

Self Cite

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

Section: Multi-modal Person Re-identificationmentioning

confidence: 99%

Section: Multi-modal Person Re-identificationmentioning

confidence: 99%

Section: Multi-modal Person Re-identificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

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

Uddin

Bhuiyan

Bappee

et al. 2023

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Nowadays deep learning models are pioneer in surveillance applications, but to achieve higher performance in reidentification persons, dissimilarity space is a better approach [22,23]. In [22] dissimilarity representation is used to identify differences occurred in source and target captured videos.…”

Section: Related Workmentioning

confidence: 99%

Dissimilarity space reinforced with manifold learning and latent space modeling for improved pattern classification

2021

View full text Add to dashboard Cite

Dissimilarity representation plays a very important role in pattern recognition due to its ability to capture structural and relational information between samples. Dissimilarity space embedding is an approach in which each sample is represented as a vector based on its dissimilarity to some other samples called prototypes. However, lack of neighborhood-preserving, fixed and usually considerable prototype set for all training samples cause low classification accuracy and high computational complexity. To address these challenges, our proposed method creates dissimilarity space considering the neighbors of each data point on the manifold. For this purpose, Locally Linear Embedding (LLE) is used as an unsupervised manifold learning algorithm. The only goal of this step is to learn the global structure and the neighborhood of data on the manifold and mapping or dimension reduction is not performed. In order to create the dissimilarity space, each sample is compared only with its prototype set including its k-nearest neighbors on the manifold using the geodesic distance metric. Geodesic distance metric is used for the structure preserving and is computed using the weighted LLE neighborhood graph. Finally, Latent Space Model (LSM), is applied to reduce the dimensions of the Euclidean latent space so that the second challenge is resolved. To evaluate the resulted representation ad so called dissimilarity space, two common classifiers namely K Nearest Neighbor (KNN) and Support Vector Machine (SVM) are applied. Experiments on different datasets which included both Euclidean and non-Euclidean spaces, demonstrate that using the proposed approach, classifiers outperform the other basic dissimilarity spaces in both accuracy and runtime.

show abstract