Applying Metric Learning: A Logo Recognition Case Study
summary
Describes applying metric learning to logo recognition at YouScan: training an embedding model once so new logo classes can be added by providing 5–10 reference images instead of retraining. Covers the task and dataset, different loss functions and approaches, evaluation metrics used, practical pitfalls and specifics, and which techniques worked or failed in their experiments.
AI-written summary · about
related talks
18mThe Unsung Hero of Vector Database: Metric Learning
40mTopological Techniques for Unsupervised Learning
41mSearching for Products by Photo
20mPragmatic Deep Learning for image labelling
46mML Use Cases at Think Big Analytics
38mPractical Few-Shot Learning
25mBrand recognition in real-life photos
43mMetric Learning for Individual Fairness