Monocular Depth Estimation Modification Using Pix2Pix Model with SELU and Alpha Dropout

Muhammad Darmawanfadilah*, Arifah Nur Ainia, Darlis Herumurti, Imam Kuswardayan

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This research focuses on modifying the Pix2Pix model for the purpose of depth estimation, which involves translating original images into RGB depth images. Depth estimation refers to the process of predicting the distance or depth information of objects in an image or scene. By incorporating the SELU activation function and employing the Alpha Dropout technique, we introduce modifications to the Pix2Pix model. The experimental results demonstrate that these modifications lead to a notable reduction in the discriminator loss by 0.09363725 and a decrease in the generator loss by 0.06176615 during the 14th iteration. These findings indicate a significant improvement in the performance of the Pix2Pix model for depth estimation after the applied modifications.

Original languageEnglish
Title of host publication2023 14th International Conference on Information and Communication Technology and System, ICTS 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages67-71
Number of pages5
ISBN (Electronic)9798350312164
DOIs
Publication statusPublished - 2023
Event14th International Conference on Information and Communication Technology and System, ICTS 2023 - Surabaya, Indonesia
Duration: 4 Oct 20235 Oct 2023

Publication series

Name2023 14th International Conference on Information and Communication Technology and System, ICTS 2023

Conference

Conference14th International Conference on Information and Communication Technology and System, ICTS 2023
Country/TerritoryIndonesia
CitySurabaya
Period4/10/235/10/23

Keywords

  • Alpha Dropout
  • GAN
  • Image translation
  • Pix2Pix
  • SELU

Fingerprint

Dive into the research topics of 'Monocular Depth Estimation Modification Using Pix2Pix Model with SELU and Alpha Dropout'. Together they form a unique fingerprint.

Cite this