Input Optimizer

Input Optimizer is a PyTorch code designed to optimize input images for any deep learning network based on a specific target. The code was developed in the Human and Robot Interaction Laboratory (Taarlab).

Optimizing Input Noise on Mark1501 Dataset

One of the library's features is the ability to optimize input noise going through a pre-trained network on Mark1501 dataset. The following images demonstrate how the library optimizes based on a specific attribute:

| |

Optimizing Input Noise on Pre-trained AlexNet Model

Another feature of Input Optimizer is the ability to optimize input noise going through a pre-trained AlexNet model. The following images demonstrate how the library optimizes based on a specific attribute:

Requirements

To use Input Optimizer, you need to have the following Python packages installed:

pythonCopy code
torch
torchvision
copy
numpy
PIL

For more information, please contact the developer at [email protected] or [email protected].