Compact Triplet Loss for person re-identification in camera sensor networks

Si, Tongzhen; Liu, Shuang

doi:10.1016/j.adhoc.2019.101984

Cited by 15 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This problem is related to pedestrian detection, but presents some important differences, and is defined as the task of identifying and matching the same individuals either across various cameras or across time within a single camera [59]. Along this line, many methods for person re-identification using triplet loss and reaching SOTA results have been proposed in the past few years, as in [60][61][62][63][64][65]. Interesting work was proposed by Wang et al [66], in which the triplet loss function is used to adjust the feature distance of each pedestrian to distinguish different pedestrians in crowded scenarios.…”

Section: Related Workmentioning

confidence: 99%

Two-Stage Pedestrian Detection Model Using a New Classification Head for Domain Generalization

Schulz,

Perez

2023

Sensors

View full text Add to dashboard Cite

Pedestrian detection based on deep learning methods have reached great success in the past few years with several possible real-world applications including autonomous driving, robotic navigation, and video surveillance. In this work, a new neural network two-stage pedestrian detector with a new custom classification head, adding the triplet loss function to the standard bounding box regression and classification losses, is presented. This aims to improve the domain generalization capabilities of existing pedestrian detectors, by explicitly maximizing inter-class distance and minimizing intra-class distance. Triplet loss is applied to the features generated by the region proposal network, aimed at clustering together pedestrian samples in the features space. We used Faster R-CNN and Cascade R-CNN with the HRNet backbone pre-trained on ImageNet, changing the standard classification head for Faster R-CNN, and changing one of the three heads for Cascade R-CNN. The best results were obtained using a progressive training pipeline, starting from a dataset that is further away from the target domain, and progressively fine-tuning on datasets closer to the target domain. We obtained state-of-the-art results, MR−2 of 9.9, 11.0, and 36.2 for the reasonable, small, and heavy subsets on the CityPersons benchmark with outstanding performance on the heavy subset, the most difficult one.

show abstract

Section: Related Workmentioning

confidence: 99%

Two-Stage Pedestrian Detection Model Using a New Classification Head for Domain Generalization

Schulz,

Perez

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Many researchers study the similarity measurement to obtain robust person Re-ID models [12], [14], [40], [41]. Hermans et al [40] employed the hardest positive and negative samples to calculate the loss value so as to generate a proper distance among pedestrian images.…”

Section: Related Work a Person Re-identificationmentioning

confidence: 99%

“…Chen et al [12] introduced the quadruplet loss that treats four pedestrian images as a unit in order to raise inter-class variations and meanwhile reduce intra-class variations. Si et al [41] introduced the Compact Triplet Loss (CTL) for person Re-ID, which makes the training samples be close to the corresponding feature centers and the different feature centers be away from each other.…”

Section: Related Work a Person Re-identificationmentioning

confidence: 99%

Semantic Constraint GAN for Person Re-Identification in Camera Sensor Networks

Liu

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel data augmentation method named Semantic Constraint Generative Adversarial Network (SCGAN) for person re-identification (Re-ID) in camera sensor networks. The proposed SCGAN can generate multiple style pedestrian images with high-level semantic information. To this end, we design two types of semantic constraints, i.e., attention constraint and identity constraint. The attention constraint aims to restrict the significant areas in the attention map to be consistent before and after image transformation. The identity constraint focuses on keeping the identity of the generated pedestrian image to be the same as that of the real one. After generating pedestrian images using SCGAN, we combine them with the real pedestrian images to train the person Re-ID model. Since the proposed SCGAN increases the diversity of training samples, the generalization of Re-ID model is enhanced. We evaluate the proposed SCGAN on three large-scale person Re-ID databases, i.e., Market1501, CUHK03 and DukeMTMC-reID, and experimental results reveal that the proposed SCGAN yields consistent improvements over other methods.

show abstract

“…[ 33 ] introduced a focal loss to the triplet criterion, assigning more weight to negative samples compared to positives. Additionally, computing centroids, such as the mean of all samples belonging to each instance, helps with cluster creations as well [ 34 , 35 , 36 , 37 ]. The triplets can be extended to a quadruplet loss, with two negative samples and one positive pair, to enlarge the inter-class variations [ 38 ] or to N-tuple loss for joint optimization of multiple instances [ 39 ].…”

Section: Introductionmentioning

confidence: 99%

A Generative Approach to Person Reidentification

Asperti,

Fiorilla,

Orsini

2024

Sensors

View full text Add to dashboard Cite

Person Re-identification is the task of recognizing comparable subjects across a network of nonoverlapping cameras. This is typically achieved by extracting from the source image a vector of characteristic features of the specific person captured by the camera. Learning a good set of robust, invariant and discriminative features is a complex task, often leveraging contrastive learning. In this article, we explore a different approach, learning the representation of an individual as the conditioning information required to generate images of the specific person starting from random noise. In this way we decouple the identity of the individual from any other information relative to a specific instance (pose, background, etc.), allowing interesting transformations from one identity to another. As generative models, we use the recent diffusion models that have already proven their sensibility to conditioning in many different contexts. The results presented in this article serve as a proof-of-concept. While our current performance on common benchmarks is lower than state-of-the-art techniques, the approach is intriguing and rich of innovative insights, suggesting a wide range of potential improvements along various lines of investigation.

show abstract

Compact Triplet Loss for person re-identification in camera sensor networks

Cited by 15 publications

References 22 publications

Two-Stage Pedestrian Detection Model Using a New Classification Head for Domain Generalization

Two-Stage Pedestrian Detection Model Using a New Classification Head for Domain Generalization

Semantic Constraint GAN for Person Re-Identification in Camera Sensor Networks

A Generative Approach to Person Reidentification

Contact Info

Product

Resources

About