Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Apr 17, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Boosting self-supervised multi-frame depth estimation with hybrid geometric-semantic constraints.

Jiaojiao Fang¹

¹School of Electronics and Information Engineering, Xi'an Jiaotong University, No. 28, Xianning West Road, Xi 'an City, Shaanxi Province, 710049, China.

Neural Networks : the Official Journal of the International Neural Network Society

|April 15, 2026

Summary

This summary is machine-generated.

Related Concept Videos

Depth Perception and Spatial Vision

Depth Perception and Spatial Vision

Depth perception is the ability to perceive objects three-dimensionally. It relies on two types of cues: binocular and monocular. Binocular cues depend on the combination of images from both eyes and how the eyes work together. Since the eyes are in slightly different positions, each eye captures a slightly different image. This disparity between images, known as binocular disparity, helps the brain interpret depth. When the brain compares these images, it determines the distance to an object.

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

ULCYP: A Multitask Model for Predicting P450 Inducers Based on Positive-Unlabeled Learning.

Journal of chemical information and modeling·2026

Same author

UGTformer: a pretrained graph transformer model for predicting UDP-glucuronosyltransferase-mediated drug metabolism.

Journal of cheminformatics·2026

Same author

Prediction of Intravenous Pharmacokinetic Parameters across Multiple Species by a Multifidelity Deep Learning Framework.

Journal of chemical information and modeling·2026

Same author

SGEDiff: a subgraph-enriched diffusion model for structure-based 3D molecular generation.

Journal of cheminformatics·2025

Same author

Computational Insights into the Regioselective Hydroxylation of Nirmatrelvir Metabolized by Cytochrome P450 3A4.

Journal of chemical information and modeling·2025

Same author

Interfacial Zn─O─Ti Sites for Efficient Photocatalytic Urea Synthesis from CO<sub>2</sub> and N<sub>2</sub>.

Angewandte Chemie (International ed. in English)·2025

Same journal

Anchor-based disentanglement framework for incremental multi-view clustering.

Neural networks : the official journal of the International Neural Network Society·2026

Same journal

Complex-valued amplitude-phase interference modeling for adversarially robust classification.

Neural networks : the official journal of the International Neural Network Society·2026

Same journal

TraNce: Type-aware hypergraph neural network with biological mediators for drug repositioning.

Neural networks : the official journal of the International Neural Network Society·2026

Same journal

Decentralized ADMM for factorization-based Low-rank matrix estimation.

Neural networks : the official journal of the International Neural Network Society·2026

Same journal

Memristive neuromorphic circuit design inspired by the neural mechanisms of conditioned fear.

Neural networks : the official journal of the International Neural Network Society·2026

Same journal

Q-learning based asynchronous Boolean control networks stabilization with data loss.

Neural networks : the official journal of the International Neural Network Society·2026

See all related articles

This study introduces a novel method for self-supervised multi-frame monocular depth estimation, enhancing accuracy in challenging conditions by integrating geometric priors and context-aware attention. The approach improves depth prediction by refining cost volumes and fusing multi-frame and single-frame cues.

Area of Science:

Computer Vision
Machine Learning
Robotics

Background:

Self-supervised multi-frame monocular depth estimation uses temporal cues for improved depth prediction.
Existing methods struggle with semantic and geometric inconsistencies in challenging regions (low-texture, occluded, dynamic).
Current approaches often rely on spatial context but neglect explicit geometric priors.

Purpose of the Study:

To develop a robust self-supervised monocular depth estimation method addressing inconsistencies in challenging environments.
To improve depth prediction accuracy by incorporating explicit geometric priors and context-aware attention.

Main Methods:

Introduced an adaptive hybrid geometric and context-aware attention module to refine cost volumes.
Aggregated local geometric and semantic information, propagating consistent costs guided by surface normals and epipolar geometry.

Keywords:

Cost volume aggregation Dynamic scenes Monocular depth estimation Self-supervised deep learning Visual attention

Related Experiment Videos

Last Updated: Apr 17, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Proposed an iterative multi-scale hybrid decoder for fusing multi-frame and single-frame cues.

Main Results:

The proposed method demonstrates robustness, accuracy, and strong generalization capabilities.
Validated across diverse datasets including outdoor driving (KITTI, Cityscapes), indoor environments (NYUv2, ScanNet), and adverse-weather conditions (DENSE).

Conclusions:

The novel attention module and hybrid decoder effectively overcome limitations of existing methods.
The approach significantly enhances self-supervised monocular depth estimation performance in challenging real-world scenarios.