6th Place Solution to Google Universal Image Embedding
Not on the map. Why?
Abstract (arXiv)
This paper presents the 6th place solution to the Google Universal Image Embedding competition on Kaggle. Our approach is based on the CLIP architecture, a powerful pre-trained model used to learn visual representation from natural language supervision. We also utilized the SubCenter ArcFace loss with dynamic margins to improve the distinctive power of class separability and embeddings. Finally, a diverse dataset has been created based on the test's set categories and the leaderboard's feedback. By carefully crafting a training scheme to enhance transfer learning, our submission scored 0.685 on the private leaderboard.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 11 of its 17 references that are in the map. First posted on arXiv 17 Oct 2022.
Builds on
11 of 17 references.
References · 11
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
- 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
- 2021Robust fine-tuning of zero-shot modelscited by 785
- 2015Deep Metric Learning via Lifted Structured Feature Embeddingcited by 449
- 2020Google Landmarks Dataset v2 -- A Large-Scale Benchmark for Instance-Level Recognition and Retrievalcited by 305
- 2017Automatic Spatially-aware Fashion Concept Discoverycited by 132
- 2020Products-10K: A Large-scale Product Recognition Datasetcited by 42
- 2022The Met Dataset: Instance-level Recognition for Artworkscited by 40
- 2020RP2K: A Large-Scale Retail Product Dataset for Fine-Grained Image Classificationcited by 27
- 2021Investigating the Vision Transformer Model for Image Retrieval Taskscited by 11
- 2020Google Landmark Recognition 2020 Competition Third Place Solutioncited by 9
Built on by
Citing papers per year: 2023: 1.
1 paper.