A new robust contrastive learning for unsupervised person re-identification

Article Engineering, Electrical & Electronic

Complementary Attention-Driven Contrastive Learning With Hard-Sample Exploring for Unsupervised Domain Adaptive Person Re-ID

Yuxuan Liu et al.

Summary: This paper proposes a complementary attention-driven contrastive learning with hard-sample exploring (CACHE) algorithm for unsupervised domain adaptive person re-identification. The algorithm improves the model's adaptability to the target domain by improving the clustering accuracy of pseudo-labels and exploring hard samples based on instance and cluster relationships.

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

Heterogeneous dual network with feature consistency for domain adaptation person re-identification

Hua Zhou et al.

Summary: This paper proposes a heterogeneous dual network (HDNet) framework to address the coupling problem between dual networks. It uses convolution with limited receptive fields and Transformer to capture local and long-range dependencies, and introduces a feature consistency loss (FCL) that does not rely on pseudo-labels to drive feature learning. Additionally, an adaptive channel mutual-aware (ACMA) module is proposed to focus on global and local information between channels. Experimental results demonstrate the competitive performance of the proposed method in person re-identification tasks.

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

Deep learning in food category recognition

Yudong Zhang et al.

Summary: Integrating artificial intelligence with food category recognition has been a field of interest for research, and it has the potential to revolutionize human interaction with food. The advancements in big data and deep learning have provided better recognition methods. This survey focuses on machine learning systems for food category recognition, including datasets, data augmentation, feature extraction, and algorithms, with a special emphasis on deep learning techniques.

INFORMATION FUSION (2023)

Add to Collection

Review Computer Science, Artificial Intelligence

Weakly supervised machine learning

Zeyu Ren et al.

Summary: Supervised learning aims to establish multiple mappings between training data and outputs through building a function or model, while weakly supervised learning is more applicable for medical image analysis due to the lack of sufficient labels. This review provides an overview of the latest progress in weakly supervised learning for medical image analysis, including incomplete, inexact, and inaccurate supervision, as well as introduces related works on different applications. Challenges and future developments of weakly supervised learning in medical image analysis are also discussed.

CAAI TRANSACTIONS ON INTELLIGENCE TECHNOLOGY (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

Domain adaptive attention-based dropout for one-shot person re-identification

Xulin Song et al.

Summary: The paper focuses on extracting universal domain-adaptive features using a domain-adaptive-attention-based-dropout (DAAD) layer, which outperforms existing techniques in improving domain adaptation performance on three re-ID datasets.

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS (2022)

Add to Collection

Article Oncology

LCDAE: Data Augmented Ensemble Framework for Lung Cancer Classification

Zeyu Ren et al.

Summary: The LCDAE framework proposed in this study overcomes the issue of overfitting in lung cancer classification tasks by applying various data augmentation techniques, and outperforms other state-of-the-art methods in terms of performance.

TECHNOLOGY IN CANCER RESEARCH & TREATMENT (2022)

Add to Collection

Article Computer Science, Artificial Intelligence

Cluster-Guided Asymmetric Contrastive Learning for Unsupervised Person Re-Identification

Mingkun Li et al.

Summary: This paper proposes a new method for unsupervised person re-identification, which utilizes clustering results to guide feature learning and achieves superior performance compared to existing methods. Experiments validate the effectiveness of the method on three benchmark datasets.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2022)

Add to Collection

Article Computer Science, Artificial Intelligence

Unsupervised person re-identification via K-reciprocal encoding and style transfer

Kun Xie et al.

Summary: This paper investigates the unsupervised person re-identification problem and proposes an effective framework through methods like style transfer, reciprocal encoding, and hard negative mining. Experimental results on three large-scale datasets demonstrate that the proposed method outperforms other unsupervised approaches, unsupervised domain adaptation methods, and semi-supervised learning methods.

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

ICE: Inter-instance Contrastive Encoding for Unsupervised Person Re-identification

Hao Chen et al.

Summary: The study introduces Interinstance Contrastive Encoding (ICE) to enhance previous class-level contrastive ReID methods by leveraging interinstance pairwise similarity scores. Experiments validate the effectiveness of the proposed unsupervised method ICE, which is competitive with supervised methods.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Refining Pseudo Labels with Clustering Consensus over Generations for Unsupervised Object Re-identification

Xiao Zhang et al.

Summary: The study introduces a pseudo label refinement strategy that leverages clustering consensus and temporal propagation to improve classification with dynamically changing classes, aiming to assist unsupervised object re-identification.

2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021 (2021)

Add to Collection

Article Computer Science, Artificial Intelligence