# Data Fine-tuning

## TL;DR

This paper introduces Data Fine-tuning, a method to enhance commercial facial analysis systems' accuracy by adding minimal perturbations to input data, without altering the models themselves.

## Contribution

It proposes a novel data perturbation approach to improve model performance without access to or modification of model parameters.

## Key findings

- Effective on LFW, CelebA, and MUCT datasets
- Improves classification accuracy without model fine-tuning
- Maintains visual appearance of input images

## Abstract

In real-world applications, commercial off-the-shelf systems are utilized for performing automated facial analysis including face recognition, emotion recognition, and attribute prediction. However, a majority of these commercial systems act as black boxes due to the inaccessibility of the model parameters which makes it challenging to fine-tune the models for specific applications. Stimulated by the advances in adversarial perturbations, this research proposes the concept of Data Fine-tuning to improve the classification accuracy of a given model without changing the parameters of the model. This is accomplished by modeling it as data (image) perturbation problem. A small amount of "noise" is added to the input with the objective of minimizing the classification loss without affecting the (visual) appearance. Experiments performed on three publicly available datasets LFW, CelebA, and MUCT, demonstrate the effectiveness of the proposed concept.