Research map / Probabilistic and scientific modeling / Variational and generative sampling
Learning Generative Vision Transformer with Energy-Based Latent Space for Saliency Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace ResNet50 encoder with Swin transformer, discriminative (ablation) | Ours-D · NJU2K | S-measure: .908 -> .919 | Helped |
| replace ResNet50 encoder with Swin transformer in generative framework (ablation) | Ours (EBM generative) · PASCAL-S | S-measure: .859 -> .877 | Helped |
| replace DPT encoder with Swin (ablation) | Ours (EBM generative) · PASCAL-S | S-measure: .870 -> .877 | Helped |
| add latent variables with EBM prior (generative learning) to Swin model (ablation) | Ours-Swin-D · DES | S-measure: .931 -> .945 | Helped |
| replace Gaussian-prior VAE with energy-based latent prior (ablation) | generative vision transformer · DES | S-measure: .942 -> .945 | Helped |
| replace Gaussian-prior ABP with energy-based latent prior (ablation) | generative vision transformer · NJU2K | S-measure: .920 -> .929 | Helped |
| replace Gaussian-prior GAN with energy-based latent prior (ablation) | generative vision transformer · NJU2K | S-measure: .928 -> .929 | Helped |
| vary EBM prior size Ce 20/60/100 (ablation) | Ours · RGB-D SOD five datasets | S-measure, F-measure | Mixed |
| vary latent dimension d from 8 to 64 (ablation) | Ours · RGB-D SOD | accuracy | No clear effect |
| vary Langevin step size within ranges (ablation) | Ours · RGB-D SOD | accuracy | No clear effect |
| proposed generative transformer vs best prior method VST | Ours · PASCAL-S | MAE: .067 -> .054 | Helped |
Closest papers by meaning
- Energy-Based Generative Cooperative Saliency Prediction (AAAI 2022 · similarity 0.90)
- LaVin-DiT: Large Vision Diffusion Transformer (CVPR 2025 · similarity 0.86)
- Learning normalized image densities via dual score matching (NeurIPS 2025 · similarity 0.85)
- Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models (CVPR 2025 · similarity 0.85)
- DiMSOD: A Diffusion-Based Framework for Multi-Modal Salient Object Detection (AAAI 2025 · similarity 0.85)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.