We propose unsupervised multi-scenario (UMS) person re-identification (ReID) as a new task that expands ReID across diverse scenarios (cross-resolution, clothing change, etc.) within a single coherent framework. To tackle UMS-ReID, we introduce image-text knowledge modeling (ITKM) -- a three-stage framework that effectively exploits the representational power of vision-language models. We start with a pre-trained CLIP model with an image encoder and a text encoder. In Stage I, we introduce a scenario embedding in the image encoder and fine-tune the encoder to adaptively leverage knowledge from multiple scenarios. In Stage II, we optimize a set of learned text embeddings to associate with pseudo-labels from Stage I and introduce a multi-scenario separation loss to increase the divergence between inter-scenario text representations. In Stage III, we first introduce cluster-level and instance-level heterogeneous matching modules to obtain reliable heterogeneous positive pairs (e.g., a visible image and an infrared image of the same person) within each scenario. Next, we propose a dynamic text representation update strategy to maintain consistency between text and image supervision signals. Experimental results across multiple scenarios demonstrate the superiority and generalizability of ITKM; it not only outperforms existing scenario-specific methods but also enhances overall performance by integrating knowledge from multiple scenarios.

Clothing change person re-identification (CC-ReID) aims to match different images of the same person, even when the clothing varies across images. To reduce manual labeling costs, existing unsupervised CC-ReID methods employ clustering algorithms to generate pseudo-labels. However, they often fail to assign the same pseudo-label to two images with the same identity but different clothing—referred to as a clothing change positive pair—thus hindering clothing-invariant feature learning. To address this issue, we propose the identity-clothing similarity modeling (ICSM) framework. To effectively connect clothing change positive pairs, ICSM first performs clothing-aware learning to leverage all discriminative information, including clothing, to obtain compact clusters. It then extracts cluster-level identity and clothing features and performs inter-cluster similarity estimation to identify clothing change positive clusters, reliable negative clusters, and hard negative clusters for each compact cluster. During optimization, we design an adaptive version of existing optimization methods to enhance similarities of clothing change positive pairs, while also introducing text semantics as a supervisory signal to further promote clothing invariance. Extensive experimental results across multiple datasets validate the effectiveness of the proposed framework, demonstrating its superiority over existing unsupervised methods and its competitiveness with some supervised approaches.

Visible-infrared person re-identification (VI-ReID) aims to match visible and infrared images of the same individual. Supervised VI-ReID (SVI-ReID) methods have achieved promising performance under the guidance of manually annotated identity labels. However, the substantial annotation cost severely limits their scalability in real-world applications. As a result, unsupervised VI-ReID (UVI-ReID) methods have attracted increasing attention. These methods typically rely on pseudo-labels generated by clustering and matching algorithms to replace manual annotations. Nevertheless, the quality of pseudo-labels is often difficult to guarantee, and low-quality pseudo-labels can significantly hinder model performance improvements. To address these challenges, we explore the use of attribute arrays extracted by a large vision-language model (LVLM) to enhance VI-ReID, and propose a novel LVLM-driven attribute-aware modeling (LVLM-AAM) approach. Specifically, we first design an attribute-aware reliable labeling strategy, which refines intra-modality clustering results based on image-level attributes and improves inter-modality matching by grouping clusters according to cluster-level attributes. Next, we develop an explicit-implicit attribute fusion module, which integrates explicit and implicit attributes to obtain more fine-grained identity-related text features. Finally, we introduce an attribute-aware contrastive learning module, which jointly leverages static and dynamic text features to promote modality-invariant feature learning. Extensive experiments conducted on VI-ReID datasets validate the effectiveness of the proposed LVLM-AAM and its individual components. LVLM-AAM not only significantly outperforms existing unsupervised methods but also surpasses several supervised methods.

Clothing change person re-identification (CC-ReID) is a crucial task in intelligent surveillance, aiming to match images of the same person wearing different clothing. Promising performance in existing CC-ReID methods is achieved at the cost of labor-intensive manual annotation of identity labels. While some researchers have explored unsupervised CC-ReID, these methods still depend on additional deep learning models for preprocessing. To eliminate the need for additional models and improve performance, we propose a joint augmentation and part learning (JAPL) framework that obtains clothing change positive pairs in an unsupervised fashion by synergistically combining augmentation-based invariant learning (AugIL) and part-based invariant learning (ParIL). AugIL first constructs clothing change pseudo-positive pairs and then encourages the model to focus on clothing-invariant information by enhancing feature consistency between the pseudo-positive pairs. ParIL beneficially encourages high similarity between inter-cluster clothing change positive pair using part images and a prediction sharpening loss. PartIL also introduces a soft consistency loss that promotes clothing-invariant feature learning by encouraging consistency of class vectors between the real features actually used for CC-ReID and the part features. Experimental results on multiple ReID datasets demonstrate that the proposed JAPL not only surpasses existing unsupervised methods but also achieves competitive performance compared to some supervised CC-ReID methods.

Cross-resolution person re-identification (CR-ReID) aims to match low-resolution (LR) and high-resolution (HR) images of the same individual. To reduce the cost of manual annotation, existing unsupervised CR-ReID methods typically rely on cross-resolution fusion to obtain pseudo-labels and resolution-invariant features. However, the fusion process requires two encoders and a fusion module, which significantly increases computational complexity and reduces efficiency. To address this issue, we propose a robust labeling and invariance modeling (RLIM) framework, which utilizes a single encoder to tackle the unsupervised CR-ReID problem. To obtain pseudo-labels robust to resolution gaps, we develop cross-resolution robust labeling (CRL), which utilizes two clustering criteria to encourage cross-resolution positive pairs to cluster together and exploit the reliable relationships between images. We also introduce random texture augmentation (TexA) to enhance the model's robustness to noisy textures related to artifacts and backgrounds by randomly adjusting texture strength. During the optimization process, we introduce the resolution-cluster consistency loss, which promotes resolution-invariant feature learning by aligning inter-resolution distances with intra-cluster distances. Experimental results on multiple datasets demonstrate that RLIM not only surpasses existing unsupervised methods, but also achieves performance close to some supervised CR-ReID methods. Code is available at https://github.com/zqpang/RLIM.

Cross-resolution person re-identification (CR-ReID) aims to match images of the same person with different resolutions in different scenarios. Existing CR-ReID methods achieve promising performance by relying on large-scale manually annotated identity labels. However, acquiring manual labels requires considerable human effort, greatly limiting the flexibility of existing CR-ReID methods. To address this issue, we propose a dual-resolution fusion modeling (DRFM) framework to tackle the CR-ReID problem in an unsupervised manner. Firstly, we design a cross-resolution pseudo-label generation (CPG) method, which initially clusters high-resolution images and then obtains reliable identity pseudo-labels by fusing class vectors in both resolution spaces. Subsequently, we develop a cross-resolution feature fusion (CRFF) module to fuse features from both high-resolution and low-resolution spaces. The fusion features have the potential to serve as a new form of resolution-invariant features. Finally, we introduce cross-resolution contrastive loss and probability sharpening loss in DRFM to facilitate resolution-invariant learning and effectively utilize ambiguous samples for optimization. Experimental results on multiple CR-ReID datasets demonstrate that the proposed DRFM not only outperforms existing unsupervised methods but also approaches the performance of early supervised methods.

Visible-infrared person re-identification (VI-ReID) is a challenging cross-modality image retrieval task. Compared to visible modality person re-identification that handles only the intra-modality discrepancy, VI-ReID suffers from an additional modality gap. Most existing VI-ReID methods achieve promising accuracy in a supervised setting, but the high annotation cost limits their scalability to real-world scenarios. Although a few unsupervised VI-ReID methods already exist, they typically rely on intra-modality initialization and cross-modality instance selection, despite the additional computational time required for intra-modality initialization. In this paper, we study the fully unsupervised VI-ReID problem and propose a novel cross-modality hierarchical clustering and refinement (CHCR) method by promoting modality-invariant feature learning and improving the reliability of pseudo-labels. Unlike conventional VI-ReID methods, CHCR does not rely on any manual identity annotation and intra-modality initialization. First, we design a simple and effective cross-modality clustering baseline that clusters between modalities. Then, to provide sufficient inter-modality positive sample pairs for modality-invariant feature learning, we propose a cross-modality hierarchical clustering algorithm to promote the clustering of inter-modality positive samples into the same cluster. In addition, we develop an inter-channel pseudo-label refinement algorithm to eliminate unreliable pseudo-labels by checking the clustering results of three channels in the visible modality. Extensive experiments demonstrate that CHCR outperforms state-of-the-art unsupervised methods and achieves performance competitive with many supervised methods.

RGB (visible), near-infrared (NI), and thermal infrared (TI) imaging modalities are commonly combined for round-the-clock surveillance. We introduce a novel unsupervised multi-modality person re-identification (MM-ReID) task, which, based on an individual’s image in any one modality, seeks to identify matches in the other two modalities. Compared to prior MM-ReID problem formulations, unsupervised MM-ReID significantly reduces labeling cost and imaging constraints. To address the unsupervised MM-ReID task, we propose a novel inter-modality similarity learning (IMSL) framework consisting of four synergistic interconnected modules: modality mean clustering (MMC), multi-modality reliability estimation (MMRE), shape-based mutual reinforcement (SMR), and modality-aware invariant learning (MIL). MMC iterates with SMR and MIL in a mutually beneficial manner to provide pseudo-labels that are robust to modality gap. MMRE normalizes sample weights, mitigating the impact of noisy labels in the multi-modality setting. SMR emphasizes shape information to implicitly enhance the model’s robustness to the modality gap and is additionally guided by pseudo-labels provided by MMC to attend to identity-related details. MIL explicitly encourages learning of modality-invariant and identity-related features via contrastive feedback for the MMC module. Extensive experimental results on the multi-modality and cross-modality datasets demonstrate that IMSL provides substantial performance gains over existing methods. Code is made available at https://github.com/zqpang/IMSL.

Abstract The core idea of active learning is to obtain higher model performance with less annotation cost. This paper proposes an independency‐enhancing adversarial active learning method. Independency‐enhancing adversarial active learning is different from the previous methods and pays more attention to sample independence. Specifically, it is believed that the informativeness of a group of samples is related to sample independence rather than the simple sum of the informativeness of each sample in the group. Therefore, an independent sample selection module based on hierarchical clustering is designed to ensure sample independence. An adversarial approach is used to learn the feature representation of a sample and use the predicted loss value to label the state of the sample. Finally, samples are selected according to the uncertainty of the samples, the diversity of the samples and the independence of the samples. The experimental results on four datasets (CIFAR‐100, Caltech‐101, Cityscapes and BDD100K) demonstrate the effectiveness and superiority of independency‐enhancing adversarial active learning.

Abstract Privacy protection in the computer vision field has attracted increasing attention. Generative adversarial network-based methods have been explored for identity anonymization, but they do not take into consideration semantic information of images, which may result in unrealistic or flawed facial results. In this paper, we propose a Semantic-aware De-identification Generative Adversarial Network (SDGAN) model for identity anonymization. To retain the facial expression effectively, we extract the facial semantic image using the edge-aware graph representation network to constraint the position, shape and relationship of generated facial key features. Then the semantic image is injected into the generator together with the randomly selected identity information for de-Identification. To ensure the generation quality and realistic-looking results, we adopt the SPADE architecture to improve the generation ability of conditional GAN. Meanwhile, we design a hybrid identity discriminator composed of an image quality analysis module, a VGG-based perceptual loss function, and a contrastive identity loss to enhance both the generation quality and ID anonymization. A comparison with the state-of-the-art baselines demonstrates that our model achieves significantly improved de-identification (De-ID) performance and provides more reliable and realistic-looking generated faces. Our code and data are available on https://github.com/kimhyeongbok/SDGAN

Fully unsupervised person re-identification (ReID) methods aim to learn discriminative features without using labeled ReID data. Because these methods are easily affected by camera discrepancies, similar studies have typically designed optimization methods to enable the model to learn camera-invariant features. However, they often ignore the impact of camera discrepancies on clustering results. Specifically, camera discrepancies will reduce the intra-class camera diversity and promote the generation of noise labels. To solve the above problems, we propose a unified unsupervised learning framework: camera invariant feature learning (CIFL) framework. First, we designed a novel DBSCAN-NN algorithm in the CIFL framework that improves the intra-class camera diversity by forcibly merging samples from different cameras. Then, we designed feature ensemble clustering that improves the accuracy of the pseudo-labels by clustering feature ensembles. In addition, we designed an optimization method for camera discrepancies: stochastic pulled loss. With the stochastic pulled loss, the ReID model is forced to learn camera-invariant features. We verified the effectiveness and generalization of CIFL on four ReID datasets (Market-1501, DukeMTMC-reID, MSMT17 and CUHK03-NP). The experimental results show that CIFL not only outperforms the existing fully unsupervised methods but also is superior to the unsupervised domain adaptation methods.

The person re-identification (ReID) method in a single-domain achieves appealing performance, but its reliance on label information greatly limits its extensibility. Therefore, the unsupervised cross-domain ReID method has received extensive attention. Its purpose is to optimize the model by using the labelled source domain and the unlabelled target domain and finally make the model well generalized in the target domain. We propose an unsupervised cross-domain ReID method based on median stable clustering (MSC) and global distance classification (GDC). Specifically, the measurement method used by MSC comprehensively considers the similarity between clusters, the number of samples in a cluster, and the combined similarity within a cluster. Different from the method based on triple loss, GDC can separate the distance distribution of positive and negative sample pairs in a global scope. In addition, considering that model performance is very sensitive to probability parameters when source domain memory is reconsolidated, we designed a dynamic memory reconsolidation (DMR) method to reduce the influence of parameters on performance. Extensive experiments on large-scale datasets (Market-1501, DukeMTMC-reID and MSMT17) demonstrate the superior performance of MSC-GDC over the state-of-the-art methods.

In the cross-domain person re-identification (ReID) method based on clustering, the performance of the model depends heavily on the quality of the information it obtains from clustering. To improve the reliability of the clustering information obtained by the model, we propose a biclustering collaborative learning (BCL) framework derived from an identity disentanglement adaptation network (IDA-Net). IDA-Net encodes the identity and style of the input image and transfers the style on the premise of maintaining identity consistency. By comparing the clustering results obtained on the same dataset before and after the transfer process, BCL can select hard samples with higher confidence for model optimization. In each iteration, we design a conditional batch hard triplet loss to optimize the two networks. Extensive experiments on large-scale datasets (Maket1501, DukeMTMC-reID and MSMT17) demonstrate the superior performance of BCL over the state-of-the-art methods.