AI Paper Atlas

Research map / Benchmarks / Oxford5k

Oxford5k leaderboard

14 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on Oxford5k. Best mAP: 96.2 (Proposed diffusion w/ late fusion, AAAI 2019).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

Oxford5k · mAP

14 papers, each with its best reported number for its own method; higher is better.

#mAP ↑MethodPaperTopic
196.2Proposed diffusion w/ late fusion
R-MAC (ResNet)
Efficient Image Retrieval via Decoupling Diffusion into Online and Offline Processing
AAAI 2019 · Performance comparison table
Fast diffusion sampling
295.8FSR rank-r, regional, 21×2,048
ResNet101
Fast Spectral Ranking for Similarity Search
CVPR 2018 · Table
Prediction and decision losses
395Regional diffusion
ResNet101, 5 regions
Efficient Diffusion on Region Manifolds: Recovering Small Objects With Compact CNN Representations
CVPR 2017 · Table 2
Autoregressive and latent generation
490DELF+FT+ATT+DIR+QELarge-Scale Image Retrieval With Attentive Deep Local Features
ICCV 2017 · Table 1
CNN features
588.9FOM MaxInter+AQE
VGG16
Efficient Object Instance Search Using Fuzzy Objects Matching
AAAI 2017 · Table 4
Open-set and few-shot detection
687.8PWA
fine-tuned ResNet101
Unsupervised Part-Based Weighting Aggregation of Deep Convolutional Features for Image Retrieval
AAAI 2018 · Table: Accuracy comparison wit
Recognition and long-tail
786.1MSB+Gram-CS, multiple scaleMultiple Saliency and Channel Sensitivity Network for Aggregated Convolutional Feature
AAAI 2019 · Table 1
CNN features
886.1Deep Retrieval (fine-tuned CNN)
ResNet101
SIFT Meets CNN: A Decade Survey of Instance Retrieval
TPAMI 2017 · Table 6
Recognition and long-tail
985.1DRH All (gDRH+lDRH+QE)Deep Region Hashing for Generic Instance Search from Images
AAAI 2018 · Table 4
CNN features
1084.1Unsupervised parts, Hungarian wpCOP, K=100, 256 dimensions
ResNet101 Landmarks
Unsupervised Part Learning for Visual Recognition
CVPR 2017 · Table 5
Recognition and long-tail
4 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.