Sort out some saliency methods (2D RGB, 2D RGB-D, 360 RGB, Video SOD) and summarize (Code and Paper).
obd. : Object Detection sod. : SOD seg. : segmentation depe. : Depth Estimation rgbd. : RGB-D 360. : 360° image suv. : survey com. : compression eval. : evaluate metrics
| NO. | Keyword | Title | Paper | Code |
|---|---|---|---|---|
| 01 | suv. |
What is a Salient Object? A Dataset and a Baseline Model for Salient Object Detection | IEEE | - |
| 02 | suv. |
Salient Object Detection: A Benchmark | IEEE | C++ & Matlab |
| NO. | Keywords | Title | Paper | Code |
|---|---|---|---|---|
| 01 | obd. |
Feature Pyramid Networks for Object Detection | CVPR | - |
| 02 | seg. |
DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs | IEEE | Tensorflow |
| 03 | - | Look around you: Saliency maps for omnidirectional images in VR applications | IEEE | - |
| 04 | eval. |
Structure-measure: A New Way to Evaluate Foreground Maps | IEEE | MatLab / Python |
| NO. | Keywords | Title | Paper | Code |
|---|---|---|---|---|
| 01 | obd. 360. |
Object Detection in Equirectangular Panorama | CVPR | - |
| 02 | 360. |
Saliency Detection in 360° Videos | ECCV | PyTorch |
| 03 | eval. |
Enhanced-alignment Measure for Binary Foreground Map Evaluation | IJCAI | MatLab |
| NO. | Keywords | Title | Paper | Code |
|---|---|---|---|---|
| 01 | sod. |
A Simple Pooling-Based Design for Real-Time Salient Object Detection | CVPR | PyTorch |
| 02 | sod. |
Cascaded Partial Decoder for Fast and Accurate Salient Object Detection | CVPR | PyTorch |
| 03 | 360. |
HorizonNet: Learning Room Layout with 1D Representation and Pano Stretch Data Augmentation | CVPR | PyTorch |
| NO. | Keywords | Title | Paper | Code |
|---|---|---|---|---|
| 01 | depe. 360. |
BiFuse: Monocular 360 Depth Estimation via Bi-Projection Fusion | CVPR | PyTorch |
| 02 | 360. |
SalBiNet360: Saliency Prediction on 360° Images with Local-Global Bifurcated Deep Network | IEEE VR | - |
| 03 | sod. 360. |
Distortion-Adaptive Salient Object Detection in 360∘ Omnidirectional Images | IEEE | Caffe |
| 04 | sod. 360. |
Stage-wise Salient Object Detection in 360° Omnidirectional Image via Object-level Semantical Saliency Ranking | IEEE | PyTorch |
| 05 | sod. 360. |
FANet: Features Adaptation Network for 360° Omnidirectional Salient Object Detection | IEEE | PyTorch |
| NO. | Keywords | Title | Paper | Code |
|---|---|---|---|---|
| 01 | rgbd. sod. |
Calibrated RGB-D Salient Object Detection | CVPR | PyTorch |
| 02 | rgbd. |
Deep RGB-D Saliency Detection with Depth-Sensitive Attention and Automatic Multi-Modal Fusion | CVPR | - |
| 03 | rgbd. |
Uncertainty Inspired RGB-D Saliency Detection | CVPR | PyTorch |
| 04 | depe. 360. |
HoHoNet: 360 Indoor Holistic Understanding with Latent Horizontal Features | CVPR | PyTorch |
| 05 | depe. 360. |
UniFuse: Unidirectional Fusion for 360$^{\circ}$ Panorama Depth Estimation | CVPR | PyTorch |
| 06 | seg. |
Fully Convolutional Networks for Panoptic Segmentation | CVPR | Detectron2 |
| 07 | 360. |
一种立体全景图像显著性检测模型 | 激光与光电子学进展 | - |
| 08 | com. 360. |
OSLO: On-the-Sphere Learning for Omnidirectional images and its application to 360-degree image compression | arxiv | - |
- 360°
- Fixation Prediction
- [Salient!360] : indoor & outdoor, 85 images. (Details)
- [Stanford360] : indoor & outdoor, 12 images. (Details)
- Salient Object Detection
- [360-SOD] : indoor & outdoor, 400 training images and 100 test images. (Details)
- [F-360iSOD] : The F-360iSOD is a small-scale 360◦ dataset with totally 107 panoramic images collected from Stanford360 [1] and Salient!360 [2] which contain 85 and 22 equirectangular images, respectively. (Details)
- [360-SSOD] : construct a novel dataset containing 1,105 360° omnidirectional images with pixel-wise saliency annotations; to thebest of our knowledge, this is the largest dataset for the 360° SOD problem.(Details)
- [ASOD60K] : To facilitate the study of panoramic video salient object detection (PV-SOD), collect ASOD60K, the first large-scale PV-SOD dataset providing professional annotations, which consists of 62,455 high-resolution (4K) video frames from 67 carefully selected 360° panoramic video sequences. 10,465 key frames are annotated with rich labels, namely, super-class, sub-class, attributes, HM data, eye fifixations, bounding boxes, object masks, and instance masks. (Details)
- (pending)
- Fixation Prediction
- 2D（pending）
- [MSRA10K] : formally named as THUS10000; 195MB: images + binary masks. Pixel accurate salient object labeling for 10000 images from MSRA dataset.
- [THUR15K] : 15000 images.
- [ECSSD] : 1000 images.
- [ Judd ] : 900 images.
- [SED1/2] : 200个灰度图像以及真实标注分割.
- [DUT-OMRON] : 5168 images.
- [DUTS] : 10,553 training images and 5,019 test images.
- [HKU-IS] : 4447个具有显着对象的像素注释的图像.
- [PASCAL-S] : 850 images.