Research map / Detection and segmentation / CNN features
Region Ranking SVM for Image Classification
Check 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 |
|---|---|---|---|
| Replace average pooling with jointly learned region ranking and score combination | VGG16 image classification · PASCAL VOC 2007 test | mAP (%): 90.3 -> 92.9 | Helped |
| Replace max pooling with jointly learned region ranking and score combination | VGG16 image classification · PASCAL VOC 2007 test | mAP (%): 91.8 -> 92.9 | Helped |
| Replace fixed-base-classifier SSD with joint learning of the region evaluator and combiner | VGG16 image classification · PASCAL VOC 2007 test | mAP (%): 91.8 -> 92.9 | Helped |
| Replace average pooling with jointly learned region ranking and score combination | VGG16 action classification · PASCAL VOC 2012 Action validation | mAP (%): 81.0 -> 84.2 | Helped |
| Replace max pooling with jointly learned region ranking and score combination | VGG16 action classification · PASCAL VOC 2012 Action validation | mAP (%): 81.5 -> 84.2 | Helped |
| Replace fixed-base-classifier SSD with joint learning of the region evaluator and combiner | VGG16 action classification · PASCAL VOC 2012 Action validation | mAP (%): 81.7 -> 84.2 | Helped |
| Use RRSVM rather than the strongest listed prior method | VOC 2012 action classification · PASCAL VOC 2012 Action test | mAP (%): 84.0 -> 85.5 | Helped |
| Replace dense VGG16 classification with RRSVM region pooling | VGG16 image classification · ILSVRC 2014 validation | top-1 error (%): 24.8 -> 22.9 | Helped |
| Replace dense VGG16 classification with RRSVM region pooling | VGG16 image classification · ILSVRC 2014 validation | top-5 error (%): 7.5 -> 6.7 | Helped |
| Replace dense VGG16 classification with RRSVM region pooling | VGG16 image classification · ILSVRC 2014 test | top-5 error (%): 6.8 -> 6.8 | No clear effect |
| Replace the spread bound on image representations with an L1-norm constraint on the region-combination vector (ablation) | RRSVM · Alternative formulation analysis | solution quality | Hurt |
Closest papers by meaning
- Learning from Few Samples: Transformation-Invariant SVMs with Composition and Locality at Multiple Scales (NeurIPS 2022 · similarity 0.84)
- Unified Locally Linear Classifiers With Diversity-Promoting Anchor Points (AAAI 2018 · similarity 0.83)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.83)
- Recognize Any Regions (NeurIPS 2024 · similarity 0.81)
- Region-Based Representations Revisited (CVPR 2024 · similarity 0.80)
- Optimal Transport Aggregation for Visual Place Recognition (CVPR 2024 · similarity 0.80)
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.