The Impact of Background Removal on Performance of Neural Networks for Fashion Image Classification and Segmentation

TL;DR

Removing backgrounds from fashion images using salient object detection can improve classification accuracy for simple models but may hinder deep neural networks due to loss of contextual information and incompatibility with training techniques.

Contribution

This study systematically evaluates the effects of background removal on fashion image classification and segmentation, highlighting its benefits for shallow models and limitations for deep networks.

Findings

Background removal improves accuracy by up to 5% on shallow networks.

Deep networks experience reduced performance due to loss of background context.

Background removal is effective for simple models but problematic for complex, regularized models.

Abstract

Fashion understanding is a hot topic in computer vision, with many applications having great business value in the market. Fashion understanding remains a difficult challenge for computer vision due to the immense diversity of garments and various scenes and backgrounds. In this work, we try removing the background from fashion images to boost data quality and increase model performance. Having fashion images of evident persons in fully visible garments, we can utilize Salient Object Detection to achieve the background removal of fashion data to our expectations. A fashion image with the background removed is claimed as the "rembg" image, contrasting with the original one in the fashion dataset. We conducted extensive comparative experiments with these two types of images on multiple aspects of model training, including model architectures, model initialization, compatibility with other training tricks and data augmentations, and target task types. Our experiments show that background removal can effectively work for fashion data in simple and shallow networks that are not susceptible to overfitting. It can improve model accuracy by up to 5% in the classification on the FashionStyle14 dataset when training models from scratch. However, background removal does not perform well in deep neural networks due to incompatibility with other regularization techniques like batch normalization, pre-trained initialization, and data augmentations introducing randomness. The loss of background pixels invalidates many existing training tricks in the model training, adding the risk of overfitting for deep models.