EUVIMEDEUVIMED entdecken
Crossref

Transparent Embedding Space for Interpretable Image Recognition

Jiaqi Wang, Huafeng Liu, Liping Jing

IEEE Transactions on Circuits and Systems for Video Technology · 2024-05 · 34 (5) · 3204-3219

DOI: 10.1109/tcsvt.2023.3314769 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems for Video Technology, 2024-05. Verfasst von Jiaqi Wang und 2 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 52 Literaturangaben. Gefördert von: National Natural Science Foundation of China, National Key Research and Development Program of China, Joint Foundation of the Ministry of Education.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

When humans explain their reasoning, such as their classification decisions, they often break down an image into parts and highlight the evidence from those parts to support the concepts they have in mind. Drawing inspiration from this cognitive process, several self-explaining models have been proposed to explain predictions by part-level concepts. However, these models can be limited by their structure and difficulty in determining the effect of individual parts on the output category. To address these challenges, we introduce a self-explaining architecture that uses a plug-in transparent embedding space (TesNet) to connect high-level input patches (e.g. feature maps or tokens) with output categories. The transparent embedding space is spanned by basis concepts and constructed on the Grassmann manifold. The basis concepts are enforced to be category-aware, and within-category concepts are orthogonal to each other, ensuring the embedding space is disentangled. To reduce concept redundancy and restore the concept space structure, we introduce two concept pruning methods and a new re-training strategy to build a slimming transparent embedding space. We verify the scalability of TesNet through experiments on deep networks such as VGG, ResNet, DenseNet, and Vision Transformer. Additionally, we design several metrics for self-explaining models to quantify interpretability and compare them with state-of-the-art self-explaining methods. Our experiments demonstrate that TesNet is much more effective for classification tasks, providing better interpretability on predictions and improving final accuracy.

Förderung

  • National Natural Science Foundation of China
  • National Key Research and Development Program of China
  • Joint Foundation of the Ministry of Education
  • Beijing Natural Science Foundation
  • Fundamental Research Funds for the Central Universities
  • Chinese Academy of Sciences

Zitieren

APA

Wang, J., Liu, H., & Jing, L. (2024). Transparent Embedding Space for Interpretable Image Recognition. IEEE Transactions on Circuits and Systems for Video Technology, 34(5), 3204-3219. https://doi.org/10.1109/tcsvt.2023.3314769

Vancouver

Wang J, Liu H, Jing L. Transparent Embedding Space for Interpretable Image Recognition. IEEE Transactions on Circuits and Systems for Video Technology. 2024;34(5):3204-3219. doi:10.1109/tcsvt.2023.3314769

BibTeX
@article{Wang2024,
  author = {Wang, Jiaqi and Liu, Huafeng and Jing, Liping},
  title = {Transparent Embedding Space for Interpretable Image Recognition},
  journal = {IEEE Transactions on Circuits and Systems for Video Technology},
  year = {2024},
  volume = {34},
  number = {5},
  pages = {3204--3219},
  doi = {10.1109/tcsvt.2023.3314769}
}
Kostenlos starten →