Segmenta\c{c}\~ao e contagem de troncos de madeira utilizando deep learning e processamento de imagens

TL;DR

This paper presents a deep learning-based method for counting wood logs in images by segmenting logs with CGANs and then counting clusters, achieving high accuracy and reliable counts.

Contribution

It introduces a novel approach combining Pix2Pix CGANs for segmentation with connected components for counting wood logs.

Findings

Segmentation accuracy exceeds 89%.

Counting accuracy over 97%.

Effective application of deep learning for wood log analysis.

Abstract

Counting objects in images is a pattern recognition problem that focuses on identifying an element to determine its incidence and is approached in the literature as Visual Object Counting (VOC). In this work, we propose a methodology to count wood logs. First, wood logs are segmented from the image background. This first segmentation step is obtained using the Pix2Pix framework that implements Conditional Generative Adversarial Networks (CGANs). Second, the clusters are counted using Connected Components. The average accuracy of the segmentation exceeds 89% while the average amount of wood logs identified based on total accounted is over 97%.