Smooth-ap: Smoothing The Path Towards Large-scale Image Retrieval
2020 · Andrew Brown, Weidi Xie, Vicky Kalogeiton, et al.
Abstract
Optimising a ranking-based metric, such as Average Precision (AP), is notoriously challenging due to the fact that it is non-differentiable, and hence cannot be optimised directly using gradient-descent methods. To this end, we introduce an objective that optimises instead a smoothed approximation of AP, coined Smooth-AP. Smooth-AP is a plug-and-play objective function that allows for end-to-end training of deep networks with a simple and elegant implementation. We also present an analysis for why directly optimising the ranking based metric of AP offers benefits over other deep metric learning losses. We apply Smooth-AP to standard retrieval benchmarks: Stanford Online products and VehicleID, and also evaluate on larger-scale datasets: INaturalist for fine-grained category retrieval, and VGGFace2 and IJB-C for face retrieval. In all cases, we improve the performance over the state-of-the-art, especially for larger-scale datasets, thus demonstrating the effectiveness and scalability of
Authors
(none)
Tags
Stats
Related papers
- Robust And Decomposable Average Precision For Image Retrieval (2021)0.00
- Optimization Of Rank Losses For Image Retrieval (2023)9.43
- Learning With Average Precision: Training Image Retrieval With A Listwise Loss (2019)18.86
- Hierarchical Average Precision Training For Pertinent Image Retrieval (2022)9.10
- Rethinking The Optimization Of Average Precision: Only Penalizing Negative Instances Before Positive Ones Is Enough (2021)11.00
- Segment Augmentation And Differentiable Ranking For Logo Retrieval (2022)0.00
- Not All Pairs Are Equal: Hierarchical Learning For Average-precision-oriented Video Retrieval (2024)7.50
- Scalable Learning Of Non-decomposable Objectives (2016)0.00