Research map / Vision-language and multimodal / Human and embodied modeling
Texture Interpolation for Probing Visual Perception
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 |
|---|---|---|---|
| use trained VGG19 weights instead of random weights (ablation) | Wasserstein texture interpolation · Natural texture interpolation | interpolation texture homogeneity | Helped |
| replace trained VGG19 with random weights (ablation) | Wasserstein texture interpolation · Natural texture interpolation | interpolation texture homogeneity | Hurt |
| replace VGG19 with a single-layer multiscale architecture (ablation) | Wasserstein texture interpolation · Natural texture interpolation | interpolation texture homogeneity | Hurt |
| use Gram loss instead of Wasserstein loss (ablation) | VGG19 texture interpolation · Natural texture interpolation | interpolation texture homogeneity | Hurt |
| use Wasserstein interpolation instead of PS interpolation (ablation) | Texture interpolation · Natural texture interpolation | perceptual smoothness | No clear effect |
| add naturalistic content while preserving the power spectrum | Spectrally matched Gaussian textures · Macaque V1 | stimulus-related information | No clear effect |
| add naturalistic content while preserving the power spectrum | Spectrally matched Gaussian textures · Macaque V4 | stimulus-related information | Helped |
| decode interpolation weight from full spike trains rather than spike counts (ablation) | Macaque V4 neural activity · Naturalistic-to-Gaussian texture interpolation | decoding performance | Helped |
| use the alternative closed-form Wasserstein interpolation rather than the weighted-loss approach (ablation) | Wasserstein texture interpolation · Texture interpolation | texture noisiness | Hurt |
Closest papers by meaning
- ImageNet-trained CNNs are not biased towards texture: Revisiting feature reliance through controlled suppression (NeurIPS 2025 · similarity 0.86)
- Generative Perception of Shape and Material from Differential Motion (NeurIPS 2025 · similarity 0.86)
- Paint-it: Text-to-Texture Synthesis via Deep Convolutional Texture Map Optimization and Physically-Based Rendering (CVPR 2024 · similarity 0.85)
- Diff-2-in-1: Bridging Generation and Dense Perception with Diffusion Models (ICLR 2025 · similarity 0.85)
- ScaleCrafter: Tuning-free Higher-Resolution Visual Generation with Diffusion Models (ICLR 2024 · similarity 0.85)
- Visual Anagrams Reveal Hidden Differences in Holistic Shape Processing Across Vision Models (NeurIPS 2025 · similarity 0.85)
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.