Implementation of the paper: A-Lamp: Adaptive Layout-Aware Multi-Patch Deep Convolutional Neural Network for Photo Aesthetic Assessment
The work is split up in 3 main parts :
- Multi-Patch Subnet :
- Adaptive Patch Selection
- Orderless Aggregation layer
- Layout Aware Subnet
In addition, is included the Exploratory Data Analysis of the AVA Dataset.
Model weights: https://drive.google.com/open?id=1Pmw0IbE9Cfy-LOq7f2wngxcZqL3dW4Xv
Work done as part of PRIM project.
Team members : Fabiola Espinoza, Guillaume Balezo & Simon Albergel