# An Empirical Evaluation Study on the Training of SDC Features for Dense Pixel Matching

## TL;DR

This paper empirically investigates various training strategies for the SDC descriptor network, focusing on hyperparameters, data selection, loss functions, and schedules to optimize dense pixel matching performance.

## Contribution

It provides a comprehensive experimental analysis of training practices for SDC, offering insights and best practices for training deep neural networks in dense matching tasks.

## Key findings

- Optimal training strategies improve SDC performance
- Multi-domain training enhances generalization
- Validated best practices for deep neural network training

## Abstract

Training a deep neural network is a non-trivial task. Not only the tuning of hyperparameters, but also the gathering and selection of training data, the design of the loss function, and the construction of training schedules is important to get the most out of a model. In this study, we perform a set of experiments all related to these issues. The model for which different training strategies are investigated is the recently presented SDC descriptor network (stacked dilated convolution). It is used to describe images on pixel-level for dense matching tasks. Our work analyzes SDC in more detail, validates some best practices for training deep neural networks, and provides insights into training with multiple domain data.