Multi-Label Image Recognition with Joint Class-Aware Map Disentangling and Label Correlation Embedding

Chen, Zhao-Min; Wei, Xiu-Shen; Jin, Xin; Guo, Yanwen

doi:10.1109/icme.2019.00113

Cited by 34 publications

(6 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, CPSD improves more on ResNet101 in comparison with ResNet101-TF and Q2L-R101. We attribute this to the fact that the more powerful ap-Methods mAP CF1 OF1 MS-CMA [You et al, 2020] 61.4 60.5 73.8 SRN [Zhu et al, 2017] 62.0 58.5 73.4 CPCL [Zhou et al, 2021a] 62.3 59.2 73.0 CADM [Chen et al, 2019b] 62 proaches are much harder to trap in model overfitting. Even so, our method still introduces a considerable improvement.…”

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

Boosting Multi-Label Image Classification with Complementary Parallel Self-Distillation

Xu¹,

Hu²,

Zhou³

et al. 2022

Preprint

View full text Add to dashboard Cite

Multi-Label Image Classification (MLIC) approaches usually exploit label correlations to achieve good performance. However, emphasizing correlation like co-occurrence may overlook discriminative features of the target itself and lead to model overfitting, thus undermining the performance. In this study, we propose a generic framework named Parallel Self-Distillation (PSD) for boosting MLIC models. PSD decomposes the original MLIC task into several simpler MLIC sub-tasks via two elaborated complementary task decomposition strategies named Co-occurrence Graph Partition (CGP) and Dis-occurrence Graph Partition (DGP). Then, the MLIC models of fewer categories are trained with these sub-tasks in parallel for respectively learning the joint patterns and the category-specific patterns of labels. Finally, knowledge distillation is leveraged to learn a compact global ensemble of full categories with these learned patterns for reconciling the label correlation exploitation and model overfitting. Extensive results on MS-COCO and NUS-WIDE datasets demonstrate that our framework can be easily plugged into many MLIC approaches and improve performances of recent state-of-the-art approaches. The explainable visual study also further validates that our method is able to learn both the category-specific and cooccurring features. The source code is released at https://github.com/Robbie-Xu/CPSD.

show abstract

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

Boosting Multi-Label Image Classification with Complementary Parallel Self-Distillation

Xu¹,

Hu²,

Zhou³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Note that in literature, OF1 and CF1 are more commonly used as compared to the other metrics to evaluate models for MLIC. We compare our method with SRN [Zhu et al, 2017], CADM [Chen et al, 2019f] , ML-GCN [Chen et al, 2019e], KSSNet [Liu et al, 2018], MS-CMA [You et al, 2020], MCAR [Gao and Zhou, 2021], SSGRL Chen et al [2019d], C-Trans [Lanchantin et al, 2021], ADD-GCN [Ye et al, 2020] , ASL [Ridnik et al, 2021], MlTr-l [Cheng et al, 2022], Swin-L [Liu et al, 2021b], CvT-w24 [Wu et al, 2021] and Q2L-CvT .…”

Section: B2 Ms-cocomentioning

confidence: 99%

Deep Dependency Networks for Multi-Label Classification

Arya¹,

Gogate²

2023

Preprint

View full text Add to dashboard Cite

We propose a simple approach which combines the strengths of probabilistic graphical models and deep learning architectures for solving the multi-label classification task, focusing specifically on image and video data. First, we show that the performance of previous approaches that combine Markov Random Fields with neural networks can be modestly improved by leveraging more powerful methods such as iterative join graph propagation, integer linear programming, and 1 regularizationbased structure learning. Then we propose a new modeling framework called deep dependency networks, which augments a dependency network, a model that is easy to train and learns more accurate dependencies but is limited to Gibbs sampling for inference, to the output layer of a neural network. We show that despite its simplicity, jointly learning this new architecture yields significant improvements in performance over the baseline neural network. In particular, our experimental evaluation on three video activity classification datasets: Charades, Textually Annotated Cooking Scenes (TACoS), and Wetlab, and three multi-label image classification datasets: MS-COCO, PASCAL VOC, and NUS-WIDE show that deep dependency networks are almost always superior to pure neural architectures that do not use dependency networks.

show abstract

“…The research direction on multilabel classification can be roughly categorized as: loss functions [28,39,53], training scheme [5,52,63], classification head [30,40]. Besides, label correlation modeling [8,9,55,57] and the utilization of region features [15,36,50] are proved to be effective for multi-label classification. In light of the challenge of annotating all groundtruth labels for an image, multi-label learning in the presence of missing labels (MLML) has also attracted much research attention [11,20,56,61].…”

Section: Related Workmentioning

confidence: 99%

Knowledge Distillation from Single to Multi Labels: an Empirical Study

Zhang¹,

Qin²,

Liu³

et al. 2023

Preprint

View full text Add to dashboard Cite

Knowledge distillation (KD) has been extensively studied in single-label image classification. However, its efficacy for multi-label classification remains relatively unexplored. In this study, we firstly investigate the effectiveness of classical KD techniques, including logit-based and feature-based methods, for multi-label classification. Our findings indicate that the logit-based method is not wellsuited for multi-label classification, as the teacher fails to provide inter-category similarity information or regularization effect on student model's training. Moreover, we observe that feature-based methods struggle to convey compact information of multiple labels simultaneously. Given these limitations, we propose that a suitable dark knowledge should incorporate class-wise information and be highly correlated with the final classification results. To address these issues, we introduce a novel distillation method based on Class Activation Maps (CAMs), which is both effective and straightforward to implement. Across a wide range of settings, CAMs-based distillation consistently outperforms other methods. Code is available at https://github.com/yzqinjacob/Distill-MLC.

show abstract

Multi-Label Image Recognition with Joint Class-Aware Map Disentangling and Label Correlation Embedding

Cited by 34 publications

References 18 publications

Boosting Multi-Label Image Classification with Complementary Parallel Self-Distillation

Boosting Multi-Label Image Classification with Complementary Parallel Self-Distillation

Deep Dependency Networks for Multi-Label Classification

Knowledge Distillation from Single to Multi Labels: an Empirical Study

Contact Info

Product

Resources

About