Research map / Architectures and efficiency / RNNs and normalization
Analyzing Classifiers: Fisher Vectors and Deep Neural Networks
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 |
|---|---|---|---|
| Use DeepNet instead of Fisher | PASCAL VOC 2007 classifiers · PASCAL VOC 2007 test | mAP: 59.99% -> 72.12% | Helped |
| Use sliding-window scoring with stride 20 instead of center-crop scoring | DeepNet · PASCAL VOC 2007 test | mAP: 72.12 mAP -> 75.9 mAP | Helped |
| Use LRP-ranked local features rather than random ordering for perturbation | Fisher classifier explanations · PASCAL VOC 2007 test | heatmap quality | Helped |
| Rely on contextual cues such as water for boat classification | Fisher · PASCAL VOC 2007, boat | robustness to boats outside water | Hurt |
| Rely on a copyright tag as a horse-class cue | Fisher · PASCAL VOC 2007, horse | prediction quality after tag removal | Hurt |
| Use object shape rather than contextual or texture cues | DeepNet · PASCAL VOC 2007, sheep | AP: 49.58% -> 74.04% | Helped |
| Use more context than DeepNet | Fisher · PASCAL VOC 2007 classes | outside-inside relevance ratio | Mixed |
Closest papers by meaning
- Neural Fisher Discriminant Analysis: Optimal Neural Network Embeddings in Polynomial Time (ICML 2022 · similarity 0.84)
- RIFD-CNN: Rotation-Invariant and Fisher Discriminative Convolutional Neural Networks for Object Detection (CVPR 2016 · similarity 0.83)
- Revisiting Discriminative vs. Generative Classifiers: Theory and Implications (ICML 2023 · similarity 0.82)
- LQF: Linear Quadratic Fine-Tuning (CVPR 2021 · similarity 0.82)
- Adversarial Fisher Vectors for Unsupervised Representation Learning (NeurIPS 2019 · similarity 0.82)
- Training-free Neural Architecture Search through Variance of Knowledge of Deep Network Weights (CVPR 2025 · similarity 0.82)
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.