Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
September 5, 2025IET Image ProcessingOpen Access

MemAttn‐CL: Unified Memory, Attention, and Contrastive Learning for Enhanced Text‐to‐Image Generation

View Full Paper
Ask AI
Bookmark
Share

Authors

MHMd. Ahsan HabibBangladesh University of Business and TechnologyMWMd. Anwar Hussen WadudBangladesh University of Engineering and TechnologyMRMohammad Motiur RahmanMawlana Bhashani Science and Technology University

Discussion

Loading...

Member takes

Implication

This method improves semantic consistency and image fidelity in text-to-image generation, indicating significant advancements.

Key Points

  • The approach enhances semantic consistency and visual quality in synthetic images by utilizing contrastive learning.
  • The DM-GAN+ATT+CL framework achieved an R-precision of 95.24 and improved image fidelity in text-to-image tasks.
  • Attention mechanisms were integrated into the two-step process, producing high-quality, photo-realistic images from text descriptions.
  • Extensive experimental results demonstrate that this method consistently outperforms several state-of-the-art models across multiple datasets.

Cite This Study

Habib et al. (2025) studied this question.

synapsesocial.com/papers/68bb3d682b87ece8dc956902https://doi.org/10.1049/ipr2.70185
View Full Paper
Ask AI
Bookmark
Share