Research map / Detection and segmentation / Open-set and few-shot detection
Recognizing Unseen Attribute-Object Pair with Generative Model
Zhixiong Nan, Yang Liu, Nanning Zheng, Song-Chun Zhu
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 |
|---|---|---|---|
| add triplet loss to encoding loss (ablation) | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 3.6 -> 11.2 | Improved |
| add triplet loss to encoding loss (ablation) | encoder-decoder attribute-object model · UT-Zappos50K | top-1 accuracy: 37.8 -> 45.5 | Improved |
| add decoding loss (autoencoder reconstruction) to encoding loss (ablation) | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 3.6 -> 17.2 | Improved |
| add decoding loss to encoding loss (ablation) | encoder-decoder attribute-object model · UT-Zappos50K | top-1 accuracy: 37.8 -> 41.3 | Improved |
| add discriminative loss to encoding loss (ablation) | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 3.6 -> 15.3 | Improved |
| add discriminative loss to encoding loss (ablation) | encoder-decoder attribute-object model · UT-Zappos50K | top-1 accuracy: 37.8 -> 37.9 | No clear effect |
| add decoding loss on top of triplet + discriminative (ablation) | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 15.7 -> 17.8 | Improved |
| use separate projection functions for attribute and object (ablation) | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 16.5 -> 17.8 | Improved |
| use separate projection functions for attribute and object (ablation) | encoder-decoder attribute-object model · UT-Zappos50K | top-1 accuracy: 47.5 -> 48.3 | Improved |
| swap ResNet-18 for deeper ResNet-50 features | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 17.8 -> 19.7 | Improved |
| swap ResNet-18 for ResNet-101 features | encoder-decoder attribute-object model · UT-Zappos50K | top-1 accuracy: 48.3 -> 51.9 | Improved |
| use ResNet-18 instead of VGG-16 features | encoder-decoder attribute-object model · MIT-States | top-1 accuracy: 15.4 -> 17.8 | Improved |
Closest papers by meaning
- Adversarial Fine-Grained Composition Learning for Unseen Attribute-Object Recognition (ICCV 2019 · similarity 0.89)
- Attributes as Operators: Factorizing Unseen Attribute-Object Compositions (ECCV 2018 · similarity 0.89)
- Disentangling Visual Embeddings for Attributes and Objects (CVPR 2022 · similarity 0.88)
- Learning to Infer Unseen Single-/ Multi-Attribute-Object Compositions With Graph Networks (TPAMI 2023 · similarity 0.88)
- GUIDED: Granular Understanding via Identification, Detection, and Discrimination for Fine-Grained Open-Vocabulary Object Detection (NeurIPS 2025 · similarity 0.84)
- Intriguing Properties of Generative Classifiers (ICLR 2024 · 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.
