EndoDepthL: Lightweight Endoscopic Monocular Depth Estimation with CNN-Transformer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Li, Yangke
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916122902659072
author Li, Yangke
author_facet Li, Yangke
contents In this study, we address the key challenges concerning the accuracy and effectiveness of depth estimation for endoscopic imaging, with a particular emphasis on real-time inference and the impact of light reflections. We propose a novel lightweight solution named EndoDepthL that integrates Convolutional Neural Networks (CNN) and Transformers to predict multi-scale depth maps. Our approach includes optimizing the network architecture, incorporating multi-scale dilated convolution, and a multi-channel attention mechanism. We also introduce a statistical confidence boundary mask to minimize the impact of reflective areas. To better evaluate the performance of monocular depth estimation in endoscopic imaging, we propose a novel complexity evaluation metric that considers network parameter size, floating-point operations, and inference frames per second. We comprehensively evaluate our proposed method and compare it with existing baseline solutions. The results demonstrate that EndoDepthL ensures depth estimation accuracy with a lightweight structure.
format Preprint
id arxiv_https___arxiv_org_abs_2308_02716
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle EndoDepthL: Lightweight Endoscopic Monocular Depth Estimation with CNN-Transformer
Li, Yangke
Computer Vision and Pattern Recognition
In this study, we address the key challenges concerning the accuracy and effectiveness of depth estimation for endoscopic imaging, with a particular emphasis on real-time inference and the impact of light reflections. We propose a novel lightweight solution named EndoDepthL that integrates Convolutional Neural Networks (CNN) and Transformers to predict multi-scale depth maps. Our approach includes optimizing the network architecture, incorporating multi-scale dilated convolution, and a multi-channel attention mechanism. We also introduce a statistical confidence boundary mask to minimize the impact of reflective areas. To better evaluate the performance of monocular depth estimation in endoscopic imaging, we propose a novel complexity evaluation metric that considers network parameter size, floating-point operations, and inference frames per second. We comprehensively evaluate our proposed method and compare it with existing baseline solutions. The results demonstrate that EndoDepthL ensures depth estimation accuracy with a lightweight structure.
title EndoDepthL: Lightweight Endoscopic Monocular Depth Estimation with CNN-Transformer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2308.02716