Negative Sampling Probability Annealing for Deep Metric Learning
Original EmbeddingNet framework is available at https://github.com/kerteszg/EmbeddingNet. Kudos to the original creators: https://github.com/RocketFlash/EmbeddingNet.
This repository contains the algorithms and materials used in experiments described in Deep Metric Learning using Negative Sampling Probability Annealing, submitted to MDPI Sensors as an Open Access research article.
Dataset
NIST SD19
To prove the discriminative ability of the proposed method by experiments, the NIST SD19 dataset was used. After highlighting and transformation, different experiments were performed using a similar backbone architecture.
Results showed, that the MDIPFL based approach achieves similar performance, despite of the significantly lower number of parameters.
References
[1] Rauf Yagfarov, Vladislav Ostankovich, Aydar Akhmetzyanov. Traffic Sign Classification Using Embedding Learning Approach for Self-driving Cars, IHIET–AI 2020
[2] Gábor Kertész. Metric Embedding Learning on Multi-Directional Projections Algorithms 13.6 (2020): 133.
[3] Gábor Kertész. Combining Negative Selection Techniques for Triplet Mining in Deep Metric Learning IEEE 10th Jubilee International Conference on Computational Cybernetics and Cyber-Medical Systems (ICCC) (2022): 155-160.