Research map / Vision-language and multimodal / Visual question answering
SEA: Evaluating Sketch Abstraction Efficiency via Element-level Commonsense Visual Question Answering
Jiho Park, Sieun Choi, Jaeyoon Seo, Minho Sohn, Yeana Kim, Jihie Kim
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 |
|---|---|---|---|
| increase sketch time budget from 4s to 32s (abstraction level) | SEA on SEVA · SEVA (6 classes) | SEA score: -0.56 -> 0.43 | Improved |
| increase sketch level from 4 to 32 | SEA on SEVA · SEVA (6 classes) | prediction probability P: 0.17 -> 0.75 | Improved |
| increase sketch level from 4 to 32 | SEA on SEVA · SEVA (6 classes) | penalty: 0.53 -> 0.18 | Improved |
| use Molmo as VQA model instead of GPT-4o (ablation) | SEA element VQA · CommonSketch | precision: 0.935 -> 0.798 | Hurt |
| use Qwen2.5-VL as open-source VQA instead of GPT-4o (ablation) | SEA element VQA · CommonSketch | accuracy: 0.855 -> 0.802 | Hurt |
| replace GPT-4o+GPT-4o with GPT-OSS 20B+Qwen2.5-VL (ablation) | SEA · Human ranking agreement | agreement %: 87.8 -> 88.0 | No clear effect |
| use SmolVLM 500M as element VQA (ablation) | SEA element VQA · CommonSketch | recall: 0.832 -> 0.346 | Hurt |
| use CommonSketch instead of QuickDraw sketches | SEA / classifier · recognizability | mean class probability: 0.29 -> 0.86 | Improved |
Closest papers by meaning
- O3SLM: Open Weight, Open Data, and Open Vocabulary Sketch-Language Model (AAAI 2026 · similarity 0.89)
- Open Vocabulary Semantic Scene Sketch Understanding (CVPR 2024 · similarity 0.88)
- Enhance Sketch Recognition’s Explainability via Semantic Component-Level Parsing (AAAI 2024 · similarity 0.87)
- How to Handle Sketch-Abstraction in Sketch-Based Image Retrieval? (CVPR 2024 · similarity 0.87)
- Visual Sketchpad: Sketching as a Visual Chain of Thought for Multimodal Language Models (NeurIPS 2024 · similarity 0.87)
- SketchFusion: Learning Universal Sketch Features through Fusing Foundation Models (CVPR 2025 · 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.
