Research map / Detection and segmentation / Open-set and few-shot detection
ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models
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 |
|---|---|---|---|
| evaluate ImageNet detectors on bias-controlled ObjectNet instead of ImageNet | AlexNet through PNASNet-5L · ObjectNet 113 overlapping classes | top-1/top-5 accuracy: 40-45% drop | Hurt |
| fine-tune last layer on 8-64 ObjectNet images per class | ResNet-152 · ObjectNet 113 overlapping classes | top-1 accuracy: 29% -> 39% (8), 45% (16), 50% (64) | Helped |
| vary background/rotation/viewpoint controls | ResNet-152 · ObjectNet | top-1 accuracy: 15% / 20% / 15% gap | Hurt |
| newer, stronger ImageNet architectures | object recognition · ObjectNet | accuracy | Helped |
Closest papers by meaning
- Contemplating Real-World Object Classification (ICLR 2021 · similarity 0.91)
- Noise or Signal: The Role of Image Backgrounds in Object Recognition (ICLR 2021 · similarity 0.90)
- Spatially Invariant Unsupervised Object Detection with Convolutional Neural Networks (AAAI 2019 · similarity 0.90)
- Large-Scale Long-Tailed Recognition in an Open World (CVPR 2019 · similarity 0.90)
- Enriching ImageNet With Human Similarity Judgments and Psychological Embeddings (CVPR 2021 · similarity 0.90)
- Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty (NeurIPS 2026 · similarity 0.89)
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.