Research map / Detection and segmentation / Open-set and few-shot detection
Deep Learning Human Mind for Automated Visual 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 |
|---|---|---|---|
| Add a 128-unit output layer after a 128-unit common LSTM (ablation) | RNN EEG classifier · ImageNet 40-class EEG test split | accuracy: 78.0% -> 82.9% | Helped |
| Use the 320–480 ms signal window instead of 40–160 ms (ablation) | RNN EEG classifier · ImageNet 40-class EEG | accuracy: 77.5% -> 84.0% | Helped |
| Replace the 320–480 ms signal window with the full 40–480 ms window (ablation) | RNN EEG classifier · ImageNet 40-class EEG | accuracy: 84.0% -> 82.9% | Hurt |
| Use the 40–320 ms window instead of 320–480 ms (ablation) | RNN EEG classifier · ImageNet 40-class EEG | accuracy: 84.0% -> 79.7% | Hurt |
| Use GoogleNet image features with k-NN regression rather than fine-tuned AlexNet (ablation) | EEG feature regressor · ImageNet images, average EEG targets | MSE: 1.86 -> 0.93 | Helped |
| Use ridge regression instead of k-NN with GoogleNet features (ablation) | EEG feature regressor · ImageNet images, average EEG targets | MSE: 0.93 -> 0.73 | Helped |
| Use random forest instead of k-NN with GoogleNet features (ablation) | EEG feature regressor · ImageNet images, average EEG targets | MSE: 0.93 -> 1.53 | Hurt |
| Average EEG features across subjects rather than selecting the best-subject feature (ablation) | GoogleNet + k-NN EEG feature regressor · ImageNet images | MSE: 4.01 -> 0.93 | Helped |
| Transfer RNN-learned brain features to image classification | Brain-driven classifier · ImageNet image test split | accuracy: 89.7% | Helped |
| Use brain-driven features instead of VGG features | Image feature extraction with SVM · Caltech-101 30-class subset | accuracy: 80.0% -> 89.7% | Helped |
| Use brain-driven features instead of GoogleNet features | Image feature extraction with SVM · Caltech-101 30-class subset | accuracy: 92.6% -> 89.7% | Hurt |
Closest papers by meaning
- Decoding Natural Images from EEG for Object Recognition (ICLR 2024 · similarity 0.85)
- Transformer brain encoders explain human high-level visual responses (NeurIPS 2025 · similarity 0.85)
- Geometry of naturalistic object representations in recurrent neural network models of working memory (NeurIPS 2024 · similarity 0.84)
- Toward Generalizing Visual Brain Decoding to Unseen Subjects (ICLR 2025 · similarity 0.84)
- L-WISE: Boosting Human Visual Category Learning Through Model-Based Image Selection and Enhancement (ICLR 2025 · similarity 0.84)
- $i$MIND: Insightful Multi-subject Invariant Neural Decoding (NeurIPS 2025 · 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.