Research map / Architectures and efficiency / CNN design
Shallow and Deep Convolutional Networks for Saliency Prediction
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 |
|---|---|---|---|
| Transfer the first three convolutional layers from VGG_CNN_M instead of initializing all layers randomly | Deep Convnet · Saliency prediction | accuracy | Helped |
| Use more layers (ablation) | Deep Convnet · Held-out validation set | accuracy | Helped |
| Use more feature maps per layer (ablation) | Deep Convnet · Held-out validation set | accuracy | Helped |
| Add dropout regularization (ablation) | Deep Convnet · Held-out validation set | accuracy | No clear effect |
| Add dropout after the first fully connected layer (ablation) | Shallow Convnet · Overfitting | overfitting | No clear effect |
| Use the Deep Convnet instead of the Shallow Convnet | Saliency prediction · MIT300 | Similarity: 0.46 -> 0.52 | Helped |
| Use the Deep Convnet instead of the Shallow Convnet | Saliency prediction · iSUN validation | AUC Judd: 0.79 -> 0.80 | Mixed |
Closest papers by meaning
- Understanding and Visualizing Deep Visual Saliency Models (CVPR 2019 · similarity 0.92)
- Image Saliency Prediction in Transformed Domain: A Deep Complex Neural Network Method (AAAI 2019 · similarity 0.90)
- A Deeper Look at Saliency: Feature Contrast, Semantics, and Beyond (CVPR 2016 · similarity 0.90)
- Deep Saliency With Encoded Low Level Distance Map and High Level Features (CVPR 2016 · similarity 0.90)
- Non-Local Deep Features for Salient Object Detection (CVPR 2017 · similarity 0.89)
- Salient Object Detection With Pyramid Attention and Salient Edges (CVPR 2019 · 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.