Filters








1,017 Hits in 5.0 sec

Robust Optical Flow Estimation in Rainy Scenes [article]

Ruoteng Li, Robby T. Tan, Loong-Fah Cheong
2017 arXiv   pre-print
Optical flow estimation in the rainy scenes is challenging due to background degradation introduced by rain streaks and rain accumulation effects in the scene.  ...  To our knowledge, this is the first optical flow method specifically dealing with rain.  ...  We have introduced a robust algorithm for optical flow estimation in rainy scenes.  ... 
arXiv:1704.05239v2 fatcat:uizytgwoxza75fjpxmrcj4jc6e

GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning [article]

Haipeng Li and Kunming Luo and Shuaicheng Liu
2021 arXiv   pre-print
Existing optical flow methods are erroneous in challenging scenes, such as fog, rain, and night because the basic optical flow assumptions such as brightness and gradient constancy are broken.  ...  Experiments show that our method outperforms the state-of-art methods in both regular and challenging scenes. Code and dataset are available at https://github.com/megvii-research/GyroFlow.  ...  As discussed in Sec. 2, RainFlow [28] is designed to estimate the optical flow under rainy scenes.  ... 
arXiv:2103.13725v2 fatcat:dr7bgxe2ujbkhjbaw3eezsvvpy

Behind-The-Scenes (BTS): Wiper-Occlusion Canceling for Advanced Driver Assistance Systems in Adverse Rain Environments

Junekyo Jhung, Shiho Kim
2021 Sensors  
This study proposes behind-the-scenes (BTS) that detects and removes wiper-occlusion in real-time image inputs under rainy weather conditions.  ...  We fine-tuned a deep learning-based optical flow model with a synthesized dataset, which was generated with pseudo-ground truth wiper masks and flows using auto-labeling with acquired real rainy images  ...  We propose a behind-the-scenes (BTS) for windshield wiper-occlusion canceling by leveraging optical flow to maintain clear visibility of images while driving under rainy weather conditions.  ... 
doi:10.3390/s21238081 pmid:34884085 fatcat:kz2jkhyezzbsfkriwff57potwq

Adversarial Scene Reconstruction and Object Detection System for Assisting Autonomous Vehicle [article]

Md Foysal Haque, Hay-Youn Lim, Dae-Seong Kang
2021 arXiv   pre-print
Also, the classifiers face difficulties in identifying the contexts of the scenes.  ...  The proposed model achieved 87.3 percent accuracy for scene reconstruction and 89.2 percent in scene understanding and detection tasks.  ...  , which determines from the optical flow vectors of multiple frames.  ... 
arXiv:2110.07716v1 fatcat:dagcna5i2bc2rbym3cufwjo3nq

A Rain Pixel Recovery Algorithm for Videos
english

Bhuvaneshwari C.M elinamath Amruta
2015 International Journal of Innovative Research in Computer and Communication Engineering  
The existing methods generally work well with light rain and relatively static scenes, but when dealing with heavier rainfall in dynamic scenes in motion, these existing methods give very poor results.  ...  Extensive simulation results show that the proposed algorithm shows a much better performance for rainy scenes with large motion than the existing algorithms.  ...  Initially, K GMM components are presumed to exist [8] in the optical flow field.  ... 
doi:10.15680/ijircce.2015.0305037 fatcat:gcjcyevuebgx7lykwc3ijknaqi

Video Deraining for Mutual Motion by Fast Bilateral Filtering on Spatiotemporal Features

2020 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
The second part of the current work proposes a novel method of processing rainy videos in different mutual motion scenarios by dynamic bilateral filtering and deep auto-encoder on time-sliced video frames  ...  The proposal of the scene categorization not only classifies the scenes based on mutual motion between camera assembly and object of interest, but it also triggers and influences the algorithm of deraining  ...  of the optical flow would vary depending on the type of mutual motion available in the scene.  ... 
doi:10.35940/ijitee.b7205.019320 fatcat:glqhfwndsfhotik6wxrakaf65a

Priors for Stereo Vision under Adverse Weather Conditions

Stefan Gehrig, Maxim Reznitskii, Nicolai Schneider, Uwe Franke, Joachim Weickert
2013 2013 IEEE International Conference on Computer Vision Workshops  
They are addressed in this work. We formulate a temporal prior and a scene prior, which we apply to SGM and Graph Cut.  ...  We also outperform the ECCV Robust Vision Challenge winner, iSGM, on this database.  ...  In our implementation, we use a sparse, robust and efficient optical flow method [28] and apply a 5x5 dilation to densify it (Figure 3 top middle).  ... 
doi:10.1109/iccvw.2013.39 dblp:conf/iccvw/GehrigRSFW13 fatcat:qqntqsv2knfcvdxlrqznqick64

Not Just Streaks: Towards Ground Truth for Single Image Deraining [article]

Yunhao Ba, Howard Zhang, Ethan Yang, Akira Suzuki, Arnold Pfahnl, Chethan Chinder Chandrappa, Celso de Melo, Suya You, Stefano Soatto, Alex Wong, Achuta Kadambi
2022 arXiv   pre-print
To learn a representation robust to rain phenomena, we propose a deep neural network that reconstructs the underlying scene by minimizing a rain-robust loss between rainy and clean images.  ...  Extensive experiments demonstrate that our model outperforms the state-of-the-art deraining methods on real rainy images under various conditions.  ...  Our approach is inspired by the recent advances in contrastive learning [6] , and we aim to distill rain-robust representations of real-world scenes by directly comparing the rainy and clean images in  ... 
arXiv:2206.10779v2 fatcat:gpkxv2cktjd4jlwaug4ayxkqbq

SHIFT: A Synthetic Driving Dataset for Continuous Multi-Task Domain Adaptation [article]

Tao Sun, Mattia Segu, Janis Postels, Yuxuan Wang, Luc Van Gool, Bernt Schiele, Federico Tombari, Fisher Yu
2022 arXiv   pre-print
In this paper, we introduce the largest multi-task synthetic dataset for autonomous driving, SHIFT.  ...  It presents discrete and continuous shifts in cloudiness, rain and fog intensity, time of day, and vehicle and pedestrian density.  ...  Overcoming catastrophic forgetting in neu- optical flow, and scene flow estimation. In Proceedings of ral networks.  ... 
arXiv:2206.08367v1 fatcat:qdajmi3tpzcsrcx7x2oo3ofsfy

PERFORMANCE METRICS IN VIDEO SURVEILLANCE SYSTEM

AZHAR, A. A. SHAFIE, M. M. RASHID
2013 Journal of Engineering Science and Technology  
The algorithms and filters that can be incorporated in tracking multiples object to solve the occluded and natural busy scenes in surveillance systems are also reviewed in this paper.  ...  One of the areas that are being actively researched is on the abilities of surveillance systems to track multiple objects over time in occluded scenes and to keep a consistent identity for each target  ...  Optical flow Optical flow [24] is a two-dimensional vector field that represents velocities and their directions in each point of the image.  ... 
doaj:ccb4ac08b96c4af8aa059b8424a1ae0e fatcat:lbpubq7zmrck5mwtxumumeicxy

A Fast and Robust Lane Detection Method Based on Semantic Segmentation and Optical Flow Estimation

Sheng Lu, Zhaojie Luo, Feng Gao, Mingjie Liu, KyungHi Chang, Changhao Piao
2021 Sensors  
In terms of lane segmentation, a robust semantic segmentation network was proposed to segment key frames and a fast and slim optical flow estimation network was used to track non-key frames.  ...  Consequently, we proposed a fast and robust lane detection method by combining a semantic segmentation network and an optical flow estimation network.  ...  Optical Flow Estimation Network and Training Method In order to speed lane segmentation up and retain robustness, we combined an optical flow estimation network with a wrapping unit based on bilinear interpolation  ... 
doi:10.3390/s21020400 pmid:33430036 fatcat:jdnd56g4dbdg7acotopc5mh4bi

Traffic Parameters Detection Using Edge and Texture

Yu Qiao, Zhongke Shi
2012 Procedia Engineering  
According to the characteristics of the traffic scene, Canny edge detection is used to get the edge information of region of interest (ROI).  ...  Through experiments taken in traffic command system of Dongguan City, the results show that this approach is accurate and performs well in real-time under different weather and illumination.  ...  The computation complexity of optical flow is not conducive to real-time processing, and the frame differencing can not detect stationary vehicles at intersection.  ... 
doi:10.1016/j.proeng.2012.01.584 fatcat:ejlicpdvhvhhtgas4yjxcmp324

Raindrop Detection and Removal from Long Range Trajectories [chapter]

Shaodi You, Robby T. Tan, Rei Kawakami, Yasuhiro Mukaigawa, Katsushi Ikeuchi
2015 Lecture Notes in Computer Science  
In rainy scenes, visibility can be degraded by raindrops which have adhered to the windscreen or camera lens.  ...  Our trajectory based video completion method not only removes the raindrops but also complete the motion field, which benefits motion estimation algorithms to possibly work in rainy scenes.  ...  Video in rainy scenes and events on the trajectories. (a) A clear day scene. (b)A scene with a thick raindrop. (c) A scene with a thin raindrop. The clear scene data is from [22] .  ... 
doi:10.1007/978-3-319-16808-1_38 fatcat:jalsvaqsnbevnirdzcy2cthgxe

DeepRoad: GAN-based Metamorphic Autonomous Driving System Testing [article]

Mengshi Zhang, Yuqun Zhang, Lingming Zhang, Cong Liu, Sarfraz Khurshid
2018 arXiv   pre-print
scenes.  ...  In this paper, we propose DeepRoad, an unsupervised framework to automatically generate large amounts of accurate driving scenes to test the consistency of DNN-based autonomous driving systems across different  ...  ing in measuring the robustness of autonomous driving systems.  ... 
arXiv:1802.02295v2 fatcat:m42xaib7xzgtrfntxypturlh54

Leveraging Synthetic Data to Learn Video Stabilization Under Adverse Conditions [article]

Abdulrahman Kerim, Washington L. S. Ramos, Leandro Soriano Marcolino, Erickson R. Nascimento, Richard Jiang
2022 arXiv   pre-print
In this paper, we propose a synthetic-aware adverse weather robust algorithm for video stabilization that does not require real data and can be trained only on synthetic data.  ...  Our results show that current approaches perform poorly in at least one weather condition, and that, even training in a small dataset with synthetic data only, we achieve the best performance in terms  ...  As anticipated, not utilizing optical flow information decreased the video stabilization quality (column No Optical Flow in Table 4 ).  ... 
arXiv:2208.12763v1 fatcat:ywegzi5vmnapxeiocm454eifku
« Previous Showing results 1 — 15 out of 1,017 results