/MLNet-Pytorch

Implementation of A Deep Multi-Level Network for Saliency Prediction in Pytorch

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

MLNet-Pytorch

Implementation of A Deep Multi-Level Network for Saliency Prediction in Pytorch

Description :

As Human, We only focus on certain part of Image which is called salient region. This Project predict Saliency map of given Image which is useful in many areas like Robotics, Image-aware editing, caption generation, fast-response systems.

For Mobile Users: https://nbviewer.jupyter.org/github/immortal3/MLNet-Pytorch/blob/master/MLNet_Pytorch.ipynb

Notebook is directly runnable to Google colab.

Note : Due to Memory Limit, some layers were frozen and output saliency maps size was reduced by half during Training.

Result :

AUC CC KL SAUC IG NSS SIM
Ours 0.771 0.553 1.142 0.619 -0.380 1.014 0.573
Original Image 1 2 3 4
Predicted Saliency Map 1 2 3 4
Ground Truth 3 3 3 3

References:

[1] Cornia, Marcella, et al. "A deep multi-level network for saliency prediction." Pattern Recognition (ICPR), 2016 23rd International Conference on. IEEE, 2016.

[2] Jiang, Ming, et al. "Salicon: Saliency in context." Proceedings of the IEEE conference on computer vision and pattern recognition. 2015.