Chosen methods of improving small object recognition with weak recognizable features

TL;DR

This paper proposes a GAN-based data augmentation method to enhance small object detection accuracy in deep learning models, addressing issues like limited samples and feature representation, demonstrated on the VOC Pascal dataset.

Contribution

It introduces a novel GAN-based augmentation approach specifically designed for small object detection, improving upon traditional augmentation techniques.

Findings

GAN augmentation increases small object detection accuracy

The method outperforms traditional augmentation strategies

Improved detection results on VOC Pascal dataset

Abstract

Many object detection models struggle with several problematic aspects of small object detection including the low number of samples, lack of diversity and low features representation. Taking into account that GANs belong to generative models class, their initial objective is to learn to mimic any data distribution. Using the proper GAN model would enable augmenting low precision data increasing their amount and diversity. This solution could potentially result in improved object detection results. Additionally, incorporating GAN-based architecture inside deep learning model can increase accuracy of small objects recognition. In this work the GAN-based method with augmentation is presented to improve small object detection on VOC Pascal dataset. The method is compared with different popular augmentation strategies like object rotations, shifts etc. The experiments are based on FasterRCNN model.