Research map / Vision-language and multimodal / Human and embodied modeling
Understanding Image Impressiveness Inspired by Instantaneous Human Perceptual Cues
Jufeng Yang, Yan Sun, Jie Liang, Yong-Liang Yang, Ming-Ming Cheng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| fuse all low-level features instead of single cues (T&C&G&S vs texture) (ablation) | impressiveness classifier · Flickr Imp | accuracy: 0.539 -> 0.734 | Improved |
| use deep features (EmoNet+CaffeNet) instead of low-level fusion (ablation) | impressiveness classifier · Flickr Imp | accuracy: 0.734 -> 0.845 | Improved |
| use deep features instead of low-level fusion (ablation) | impressiveness classifier · Trip Imp | accuracy: 0.728 -> 0.808 | Improved |
| add low-level and mid-level cues to deep features (all fusion) (ablation) | impressiveness classifier · Flickr Imp | accuracy: 0.845 -> 0.852 | Improved |
| add low-level and mid-level cues to deep features (all fusion) (ablation) | impressiveness classifier · Trip Imp | accuracy: 0.808 -> 0.820 | Improved |
| use unusualness cue (C&G&S) instead of aesthetics (T&C&G) (ablation) | impressiveness classifier · Flickr Imp | accuracy: 0.660 -> 0.715 | Improved |
| add mid-level emotion and object features to low-level in recommendation (impressiveness vs low-level) | image recommendation · News Imp Sport | MAP: 0.603 -> 0.763 | Improved |
| impressiveness vs low-level features in recommendation | image recommendation · News Imp Technology | MRR: 0.768 -> 0.759 | Hurt |
| impressiveness representation vs SIFT&BoW | image recommendation · News Imp Politics | MAP: 0.664 -> 0.807 | Improved |
Closest papers by meaning
- Visual Interestingness Prediction: A Benchmark Framework and Literature Review (IJCV 2021 · similarity 0.88)
- Visual Interestingness Decoded: How GPT-4o Mirrors Human Interests (ICCV 2025 · similarity 0.86)
- The Many Senses of Visual Similarity: A Text-Prompted Image Perceptual Metric (NeurIPS 2026 · similarity 0.84)
- Is Your Image a Good Storyteller? (AAAI 2025 · similarity 0.83)
- UniAR: A Unified model for predicting human Attention and Responses on visual content (NeurIPS 2024 · similarity 0.83)
- PerceptionRubrics: Calibrating Multimodal Evaluation to Human Perception (ICML 2026 · similarity 0.83)
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.
