Research map / Image generation / GANs
Neural Photo Editing with Introspective Adversarial Networks
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 |
|---|---|---|---|
| Add MDC blocks and Orthogonal Regularization (ablation) | DenseNet · CIFAR-100+ test | error rate: 26.71 -> 25.39 | Helped |
| Replace 3x3 filters with 3d3 MDC (ablation) | DenseNet · CIFAR-100+ test | error rate: 26.71 -> 26.02 | Helped |
| Add Orthogonal Regularization (ablation) | DenseNet · CIFAR-100+ test | error rate: 26.71 -> 26.46; 26.71 -> 26.51 | Helped |
| Replace 3x3 filters with 7x7 filters (ablation) | DenseNet · CIFAR-100+ test | error rate: 26.71 -> 26.39 | Helped |
| Use Full MDC (ablation) | DenseNet · CIFAR-100+ test | error rate: 24.42 -> 23.30 | Helped |
| Add MDC, Orthogonal Regularization, and ternary adversarial loss (ablation) | IAN · CelebA | Pixel: 0.295 -> 0.192 | Helped |
| Add MDC, Orthogonal Regularization, and ternary adversarial loss (ablation) | IAN · CelebA | Inception: $1389 (\pm 64)$ -> $3627 (\pm 146)$ | Helped |
| Add MDC blocks (ablation) | IAN · CelebA | Pixel: 0.285 -> 0.248 | Helped |
| Add Orthogonal Regularization (ablation) | IAN · CelebA | Pixel: 0.285 -> 0.258 | Helped |
| Add ternary adversarial loss (ablation) | IAN · CelebA | Pixel: 0.285 -> 0.254 | Helped |
| Add Orthogonal Regularization (ablation) | DenseNet · CelebA attribute classification | validation error: 6.55\% to 4.22\% | Helped |
| Use IAN with L2-SVM instead of ALI with L2-SVM | Semi-supervised SVHN classification · SVHN, 1000 training examples | error rate: $19.14\% (\pm 0.50\%)$ -> $18.50\% (\pm 0.38\%)$ | Hurt |
Closest papers by meaning
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.90)
- IntroVAE: Introspective Variational Autoencoders for Photographic Image Synthesis (NeurIPS 2018 · similarity 0.90)
- Navigating the GAN Parameter Space for Semantic Image Editing (CVPR 2021 · similarity 0.89)
- Large Scale GAN Training for High Fidelity Natural Image Synthesis (ICLR 2019 · similarity 0.89)
- SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations (ICLR 2022 · similarity 0.89)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 2024 · similarity 0.87)
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.