REVIEW 5 cited by
Unicom: Universal and Compact Representation Learning for Image Retrieval
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
Modern image retrieval methods typically rely on fine-tuning pre-trained encoders to extract image-level descriptors. However, the most widely used models are pre-trained on ImageNet-1K with limited classes. The pre-trained feature representation is therefore not universal enough to generalize well to the diverse open-world classes. In this paper, we first cluster the large-scale LAION400M into one million pseudo classes based on the joint textual and visual features extracted by the CLIP model. Due to the confusion of label granularity, the automatically clustered dataset inevitably contains heavy inter-class conflict. To alleviate such conflict, we randomly select partial inter-class prototypes to construct the margin-based softmax loss. To further enhance the low-dimensional feature representation, we randomly select partial feature dimensions when calculating the similarities between embeddings and class-wise prototypes. The dual random partial selections are with respect to the class dimension and the feature dimension of the prototype matrix, making the classification conflict-robust and the feature embedding compact. Our method significantly outperforms state-of-the-art unsupervised and supervised image retrieval approaches on multiple benchmarks. The code and pre-trained models are released to facilitate future research https://github.com/deepglint/unicom.
Forward citations
Cited by 5 Pith papers
-
Illuminating Visual Identity in Universal Multimodal Embeddings
By adding identity-aware sampling and a contrastive loss on a new 28-dataset benchmark, the authors build multimodal embeddings that are far better at visual identity matching without losing general retrieval accuracy.
-
Gradient-Attention Guided Dual-Masking Synergetic Framework for Robust Text-based Person Retrieval
GA-DMS with the WebPerson dataset sets new state-of-the-art Rank-1 accuracy on CUHK-PEDES, ICFG-PEDES, and RSTPReid.
-
Locality-Sensitive Hashing for Efficient Hard Negative Sampling in Contrastive Learning
Binary LSH codes found by Hamming distance provide hard negatives for supervised contrastive learning at a fraction of the compute cost of exact pre-epoch sampling, with comparable or better accuracy on six benchmarks.
-
Enhancing Sample Utilization in Noise-Robust Deep Metric Learning With Subgroup-Based Positive-Pair Selection
A subgroup-based positive-pair selection framework lets noisy samples contribute to metric learning training instead of being discarded, improving retrieval and face recognition accuracy under label noise.
-
A Few-Shot Metric Learning Method with Dual-Channel Attention for Cross-Modal Same-Neuron Identification
A dual-channel attention metric learning model with Circle Loss and hard sample mining matches single neurons across two-photon and fMOST images using only 190 training pairs, claiming 77.4% recall at 90.1% specificity.
Discussion (0). Continue with ORCID to comment.