# On the Unreasonable Effectiveness of Centroids in Image Retrieval arXiv: https://arxiv.org/abs/2104.13643 · HTML: https://mlsanity.com/paper/2104.13643/ ## Key facts - Cited by 45 papers in the map (data through 14 Oct 2026). - Builds on 12 of its 21 references that are in the map. - First posted on arXiv 28 Apr 2021. ## Builds on (most cited first, 12 of 12) - [Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking](https://mlsanity.com/paper/1609.01775/) (2016) - [Omni-Scale Feature Learning for Person Re-Identification](https://mlsanity.com/paper/1905.00953/) (2019) - [Large-Scale Image Retrieval with Attentive Deep Local Features](https://mlsanity.com/paper/1612.06321/) (2016) - [A Strong Baseline and Batch Normalization Neck for Deep Person Re-identification](https://mlsanity.com/paper/1906.08332/) (2019) - [Similarity Reasoning and Filtration for Image-Text Matching](https://mlsanity.com/paper/2101.01368/) (2021) - [Spatial-Temporal Person Re-identification](https://mlsanity.com/paper/1812.03282/) (2018) - [Learning Deep Feature Representations with Domain Guided Dropout for Person Re-identification](https://mlsanity.com/paper/1604.07528/) (2016) - [In Defense of the Triplet Loss Again: Learning Robust Person Re-Identification with Fast Approximated Triplet Loss and Label Distillation](https://mlsanity.com/paper/1912.07863/) (2019) - [A Theoretically Sound Upper Bound on the Triplet Loss for Improving the Efficiency of Deep Distance Metric Learning](https://mlsanity.com/paper/1904.08720/) (2019) - [Combination of Multiple Global Descriptors for Image Retrieval](https://mlsanity.com/paper/1903.10663/) (2019) - [Centroid-based deep metric learning for speaker recognition](https://mlsanity.com/paper/1902.02375/) (2019) - [A Strong Baseline for Fashion Retrieval with Person Re-Identification Models](https://mlsanity.com/paper/2003.04094/) (2020) ## Built on by (most cited first, 30 of 45) - [An Outlook into the Future of Egocentric Vision](https://mlsanity.com/paper/2308.07123/) (2023) - [Ancilia: Scalable Intelligent Video Surveillance for the Artificial Intelligence of Things](https://mlsanity.com/paper/2301.03561/) (2023) - [DIVOTrack: A Novel Dataset and Baseline Method for Cross-View Multi-Object Tracking in DIVerse Open Scenes](https://mlsanity.com/paper/2302.07676/) (2023) - [Physical Adversarial Attacks for Surveillance: A Survey](https://mlsanity.com/paper/2305.01074/) (2023) - [TwinLiteNet: An Efficient and Lightweight Model for Driveable Area and Lane Segmentation in Self-Driving Cars](https://mlsanity.com/paper/2307.10705/) (2023) - [ShARc: Shape and Appearance Recognition for Person Identification In-the-wild](https://mlsanity.com/paper/2310.15946/) (2023) - [A General Framework for Jersey Number Recognition in Sports Video](https://mlsanity.com/paper/2405.13896/) (2024) - [GEFF: Improving Any Clothes-Changing Person ReID Model using Gallery Enrichment with Face Features](https://mlsanity.com/paper/2211.13807/) (2022) - [boldsymbolλ-Orthogonality Regularization for Compatible Representation Learning](https://mlsanity.com/paper/2509.16664/) (2025) - [CLIP-ReIdent: Contrastive Training for Player Re-Identification](https://mlsanity.com/paper/2303.11855/) (2023) - [Whole-body Detection, Recognition and Identification at Altitude and Range](https://mlsanity.com/paper/2311.05725/) (2023) - [Sports Re-ID: Improving Re-Identification Of Players In Broadcast Videos Of Team Sports](https://mlsanity.com/paper/2206.02373/) (2022) - [Enhancing person re-identification via Uncertainty Feature Fusion Method and Auto-weighted Measure Combination](https://mlsanity.com/paper/2405.01101/) (2024) - [Pose-dIVE: Pose-Diversified Augmentation with Diffusion Model for Person Re-Identification](https://mlsanity.com/paper/2406.16042/) (2024) - [Teaching What You Should Teach: A Data-Based Distillation Method](https://mlsanity.com/paper/2212.05422/) (2022) - [Uncertainty Aware Multitask Pyramid Vision Transformer For UAV-Based Object Re-Identification](https://mlsanity.com/paper/2209.08686/) (2022) - [FRIDA: Fisheye Re-Identification Dataset with Annotations](https://mlsanity.com/paper/2210.01582/) (2022) - [A Gated Attention Transformer for Multi-Person Pose Tracking](https://mlsanity.com/paper/2306.05807/) (2023) - [Learning Embeddings with Centroid Triplet Loss for Object Identification in Robotic Grasping](https://mlsanity.com/paper/2404.06277/) (2024) - [Product Re-identification System in Fully Automated Defect Detection](https://mlsanity.com/paper/2112.10324/) (2021) - [Motion-Aware Transformer For Occluded Person Re-identification](https://mlsanity.com/paper/2202.04243/) (2022) - [Combining human parsing with analytical feature extraction and ranking schemes for high-generalization person reidentification](https://mlsanity.com/paper/2207.14243/) (2022) - [MMFL-Net: Multi-scale and Multi-granularity Feature Learning for Cross-domain Fashion Retrieval](https://mlsanity.com/paper/2210.15128/) (2022) - [GMT: Effective Global Framework for Multi-Camera Multi-Target Tracking](https://mlsanity.com/paper/2407.01007/) (2024) - [A Re-ranking Method using K-nearest Weighted Fusion for Person Re-identification](https://mlsanity.com/paper/2509.04050/) (2025) - [Learning an Ensemble of Deep Fingerprint Representations](https://mlsanity.com/paper/2209.02425/) (2022) - [Fashion Image Retrieval with Multi-Granular Alignment](https://mlsanity.com/paper/2302.08902/) (2023) - [Runner re-identification from single-view running video in the open-world setting](https://mlsanity.com/paper/2310.11700/) (2023) - [3D Semantic MapNet: Building Maps for Multi-Object Re-Identification in 3D](https://mlsanity.com/paper/2403.13190/) (2024) - [CLOVER: Context-aware Long-term Object Viewpoint- and Environment- Invariant Representation Learning](https://mlsanity.com/paper/2407.09718/) (2024) ## Abstract (arXiv) Image retrieval task consists of finding similar images to a query image from a set of gallery (database) images. Such systems are used in various applications e.g. person re-identification (ReID) or visual product search. Despite active development of retrieval models it still remains a challenging task mainly due to large intra-class variance caused by changes in view angle, lighting, background clutter or occlusion, while inter-class variance may be relatively low. A large portion of current research focuses on creating more robust features and modifying objective functions, usually based on Triplet Loss. Some works experiment with using centroid/proxy representation of a class to alleviate problems with computing speed and hard samples mining used with Triplet Loss. However, these approaches are used for training alone and discarded during the retrieval stage. In this paper we propose to use the mean centroid representation both during training and retrieval. Such an aggregated representation is more robust to outliers and assures more stable features. As each class is represented by a single embedding - the class centroid - both retrieval time and storage requirements are reduced significantly. Aggregating multiple embeddings results in a significant reduction of the search space due to lowering the number of candidate target vectors, which makes the method especially suitable for production deployments. Comprehensive experiments conducted on two ReID and Fashion Retrieval datasets demonstrate effectiveness of our method, which outperforms the current state-of-the-art. We propose centroid training and retrieval as a viable method for both Fashion Retrieval and ReID applications.