AI Paper Atlas

Research map / Detection and segmentation / CNN features

Bilinear Convolutional Neural Networks for Fine-Grained Visual Recognition

Tsung‐Yu Lin, Aruni RoyChowdhury, Subhransu Maji

TPAMI 2017 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 411 citations (OpenAlex)

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 changedWhereMeasuredEffect
add domain-specific fine-tuning to NetVLADNetVLAD (VGG-D) · CUB-200-2011accuracy: 77.9 -> 81.9Improved
fine-tune B-CNN (VGG-D) end to endB-CNN (VGG-D) · Stanford Carsaccuracy: 82.9 -> 90.6Improved
replace indirectly fine-tuned FV with directly fine-tuned NetFVFV (VGG-D) · CUB-200-2011accuracy: 74.7 -> 79.9Improved
replace fully connected pooling with bilinear pooling (f1 augmentation)VGG-M trained from scratch · ImageNet valtop-1 error: 46.4 -> 38.7Improved
add f25 translation jitter to FC pooling (ablation)VGG-M FC · ImageNet valtop-1 error: 46.4 -> 39.6Improved
add f25 jitter to bilinear pooling (ablation)B-CNN VGG-M · ImageNet valtop-1 error: 38.7 -> 36.6Improved
add multi-scale features (ablation)B-CNN (VGG-D) · DTDmean per-class accuracy: 69.6 -> 72.9Improved
add multi-scale features (ablation)B-CNN (VGG-D) · MIT indoormean per-class accuracy: 72.8 -> 79.0Improved
replace FV pooling with B-CNN pooling for textureVGG-D relu5_3 features · DTD, FMD, KTH-T2b, MIT indoormean per-class accuracy: 73.6 vs 72.9 on DTD; 80.8 vs 81.6 on FMDNo clear effect
fine-tune B-CNNB-CNN (VGG-D) · MIT indooraccuracy: 72.8 -> 73.8Improved
project one feature with PCA to reduce bilinear dimension 16x (ablation)B-CNN (VGG-M) · CUB-200-2011, FGVC-Aircraftaccuracy: less than 1% loss in accuracy with 32x512No clear effect
use NetVLAD with 32 components instead of projected B-CNN of the same sizeNetVLAD · fine-grained datasetsaccuracy: about 3-4% less accurateHurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy