Yazılar Qeydlər Tədqiqat Tədris

Representation space evaluation of feature extractors for model transferability in fine-grained recognition

Ismayil Shahaliyev, Jamaladdin Hasanov

Signal, Image and Video Processing 20 , 439 (2026) Open access

• Paper
• PDF
• Code
• BibTeX
• Abstract

This paper introduces an evaluation pipeline for transferability estimation in fine-grained visual recognition tasks, with a focus on hand gesture recognition. The proposed framework provides a resource-efficient and interpretable tool for validating model suitability to downstream tasks without exhaustive fine-tuning. By analyzing the representation spaces of both convolutional and transformer-based vision models through dimensionality reduction and clustering, our pipeline incorporates a similarity-aware scoring technique that explicitly considers both intra-class variability and inter-class similarity. In experiments, our framework consistently ranked the vision transformer with advanced pre-training highest, while assigning lower importance to the standard vision transformer and convolutional models. These results indicate that the proposed approach can effectively highlight models best suited for fine-grained recognition.

@article{representation_mte2026,
  title = {Representation space evaluation of feature extractors for model transferability in fine-grained recognition},
  author = {Shahaliyev, Ismayil and Hasanov, Jamaladdin},
  journal = {Signal, Image and Video Processing},
  volume = {20},
  pages = {439},
  year = {2026},
  doi = {10.1007/s11760-026-05474-4},
  url = {https://doi.org/10.1007/s11760-026-05474-4}
}