English
Related papers

Related papers: Motion Consistency Loss for Monocular Visual Odome…

200 papers

This paper studies monocular visual odometry (VO) problem. Most of existing VO algorithms are developed under a standard pipeline including feature extraction, feature matching, motion estimation, local optimisation, etc. Although some of…

Computer Vision and Pattern Recognition · Computer Science 2017-09-26 Sen Wang , Ronald Clark , Hongkai Wen , Niki Trigoni

Multi-view geometry-based methods dominate the last few decades in monocular Visual Odometry for their superior performance, while they have been vulnerable to dynamic and low-texture scenes. More importantly, monocular methods suffer from…

Computer Vision and Pattern Recognition · Computer Science 2021-03-02 Huangying Zhan , Chamara Saroj Weerasekera , Jia-Wang Bian , Ravi Garg , Ian Reid

Most learning-based methods estimate ego-motion by utilizing visual sensors, which suffer from dramatic lighting variations and textureless scenarios. In this paper, we incorporate sparse but accurate depth measurements obtained from lidars…

Computer Vision and Pattern Recognition · Computer Science 2021-01-06 Bin Li , Mu Hu , Shuling Wang , Lianghao Wang , Xiaojin Gong

Depth estimation and 3D object detection are critical for scene understanding but remain challenging to perform with a single image due to the loss of 3D information during image capture. Recent models using deep neural networks have…

Computer Vision and Pattern Recognition · Computer Science 2019-04-19 Julie Chang , Gordon Wetzstein

Learning single image depth estimation model from monocular video sequence is a very challenging problem. In this paper, we propose a novel training loss which enables us to include more images for supervision during the training process.…

Computer Vision and Pattern Recognition · Computer Science 2023-10-27 Zhenwei Luo

Video segmentation approaches are of great importance for numerous vision tasks especially in video manipulation for entertainment. Due to the challenges associated with acquiring high-quality per-frame segmentation annotations and large…

Computer Vision and Pattern Recognition · Computer Science 2020-10-27 Mingyang Qian , Yi Fu , Xiao Tan , Yingying Li , Jinqing Qi , Huchuan Lu , Shilei Wen , Errui Ding

Learning based methods have shown very promising results for the task of depth estimation in single images. However, most existing approaches treat depth prediction as a supervised regression problem and as a result, require vast quantities…

Computer Vision and Pattern Recognition · Computer Science 2017-04-14 Clément Godard , Oisin Mac Aodha , Gabriel J. Brostow

Motion blur is one of the major challenges remaining for visual odometry methods. In low-light conditions where longer exposure times are necessary, motion blur can appear even for relatively slow camera motions. In this paper we present a…

Computer Vision and Pattern Recognition · Computer Science 2021-03-26 Peidong Liu , Xingxing Zuo , Viktor Larsson , Marc Pollefeys

Forecasting motion of a specific target object is a common problem for surgical interventions, e.g. for localization of a target region, guidance for surgical interventions, or motion compensation. Optical coherence tomography (OCT) is an…

Image and Video Processing · Electrical Eng. & Systems 2020-06-02 Marcel Bengs , Nils Gessert , Alexander Schlaefer

Monocular 3D object detection poses a significant challenge due to the lack of depth information in RGB images. Many existing methods strive to enhance the object depth estimation performance by allocating additional parameters for object…

Computer Vision and Pattern Recognition · Computer Science 2024-01-03 Wonhyeok Choi , Mingyu Shin , Sunghoon Im

In this work we present a monocular visual odometry (VO) algorithm which leverages geometry-based methods and deep learning. Most existing VO/SLAM systems with superior performance are based on geometry and have to be carefully designed for…

Computer Vision and Pattern Recognition · Computer Science 2020-02-19 Huangying Zhan , Chamara Saroj Weerasekera , Jiawang Bian , Ian Reid

Recent work has shown that CNN-based depth and ego-motion estimators can be learned using unlabelled monocular videos. However, the performance is limited by unidentified moving objects that violate the underlying static scene assumption in…

Computer Vision and Pattern Recognition · Computer Science 2019-10-04 Jia-Wang Bian , Zhichao Li , Naiyan Wang , Huangying Zhan , Chunhua Shen , Ming-Ming Cheng , Ian Reid

The problem of Multiple Object Tracking (MOT) consists in following the trajectory of different objects in a sequence, usually a video. In recent years, with the rise of Deep Learning, the algorithms that provide a solution to this problem…

Computer Vision and Pattern Recognition · Computer Science 2019-11-21 Gioele Ciaparrone , Francisco Luque Sánchez , Siham Tabik , Luigi Troiano , Roberto Tagliaferri , Francisco Herrera

We present a self-supervised learning framework to estimate the individual object motion and monocular depth from video. We model the object motion as a 6 degree-of-freedom rigid-body transformation. The instance segmentation mask is…

Computer Vision and Pattern Recognition · Computer Science 2020-05-14 Qi Dai , Vaishakh Patil , Simon Hecker , Dengxin Dai , Luc Van Gool , Konrad Schindler

Accurate fall detection for the assistance of older people is crucial to reduce incidents of deaths or injuries due to falls. Meanwhile, a vision-based fall detection system has shown some significant results to detect falls. Still,…

Computer Vision and Pattern Recognition · Computer Science 2021-04-14 Sagar Chhetri , Abeer Alsadoon , Thair Al Dala in , P. W. C. Prasad , Tarik A. Rashid , Angelika Maag

Heatmap regression based face alignment has achieved prominent performance on static images. However, the stability and accuracy are remarkably discounted when applying the existing methods on dynamic videos. We attribute the degradation to…

Computer Vision and Pattern Recognition · Computer Science 2020-11-16 Xu Sun , Zhenfeng Fan , Zihao Zhang , Yingjie Guo , Shihong Xia

This paper overviews different pose representations and metric functions in visual odometry (VO) networks. The performance of VO networks heavily relies on how their architecture encodes the information. The choice of pose representation…

Computer Vision and Pattern Recognition · Computer Science 2024-01-12 Olaya Álvarez-Tuñón , Yury Brodskiy , Erdal Kayacan

The modern image search system requires semantic understanding of image, and a key yet under-addressed problem is to learn a good metric for measuring the similarity between images. While deep metric learning has yielded impressive…

Computer Vision and Pattern Recognition · Computer Science 2017-08-08 Jian Wang , Feng Zhou , Shilei Wen , Xiao Liu , Yuanqing Lin

Visual odometry networks commonly use pretrained optical flow networks in order to derive the ego-motion between consecutive frames. The features extracted by these networks represent the motion of all the pixels between frames. However,…

Computer Vision and Pattern Recognition · Computer Science 2020-11-18 Hamed Damirchi , Rooholla Khorrambakht , Hamid D. Taghirad

The problem of tracking self-motion as well as motion of objects in the scene using information from a camera is known as multi-body visual odometry and is a challenging task. This paper proposes a robust solution to achieve accurate…

Robotics · Computer Science 2020-07-29 Jun Zhang , Mina Henein , Robert Mahony , Viorela Ila