Generalized dynamic object removal for dense stereo vision based scene mapping using synthesised optical flow

Hamilton, Oliver; Breckon, Toby P.

doi:10.1109/icip.2016.7532998

Cited by 14 publications

(14 citation statements)

References 25 publications

(42 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…State of the art solutions for estimation of dynamics either rely on accurate depth data acquired with time-of-flight sensors [35], or consider the dynamics estimation problem only for removing such regions from the reconstruction [36]. To cope with the contradicting requirements, we exploit the color consistency between consecutive frames.…”

Section: Stereo Based Reconstruction and Segmentation In Outdoor Ementioning

confidence: 99%

Stereo Vision Based Sensory Substitution for the Visually Impaired

Caraiman

Zvoristeanu

Burlacu

et al. 2019

Sensors

View full text Add to dashboard Cite

The development of computer vision based systems dedicated to help visually impaired people to perceive the environment, to orientate and navigate has been the main research subject of many works in the recent years. A significant ensemble of resources has been employed to support the development of sensory substitution devices (SSDs) and electronic travel aids for the rehabilitation of the visually impaired. The Sound of Vision (SoV) project used a comprehensive approach to develop such an SSD, tackling all the challenging aspects that so far restrained the large scale adoption of such systems by the intended audience: Wearability, real-time operation, pervasiveness, usability, cost. This article is set to present the artificial vision based component of the SoV SSD that performs the scene reconstruction and segmentation in outdoor environments. In contrast with the indoor use case, where the system acquires depth input from a structured light camera, in outdoors SoV relies on stereo vision to detect the elements of interest and provide an audio and/or haptic representation of the environment to the user. Our stereo-based method is designed to work with wearable acquisition devices and still provide a real-time, reliable description of the scene in the context of unreliable depth input from the stereo correspondence and of the complex 6 DOF motion of the head-worn camera. We quantitatively evaluate our approach on a custom benchmarking dataset acquired with SoV cameras and provide the highlights of the usability evaluation with visually impaired users.

show abstract

Section: Stereo Based Reconstruction and Segmentation In Outdoor Ementioning

confidence: 99%

Stereo Vision Based Sensory Substitution for the Visually Impaired

Caraiman

Zvoristeanu

Burlacu

et al. 2019

Sensors

View full text Add to dashboard Cite

show abstract

“…As 3D imagery has become the staple requirement within many computer vision applications, accurate and efficient depth estimation is now one of its core foundations. Conventional depth estimation methods have relied on numerous strategies such as stereo correspondence [67,28], structure from motion [14,9], depth from shading and light diffusion [73,82,1] and alike. However, these approaches are often rife with issues such as depth inhomogeneity, missing depth (holes), computationally intensive requirements and more importantly, careful calibration and setup demanding expert knowledge which often requires special post-processing [4,2,49,58].…”

Section: Introductionmentioning

confidence: 99%

Real-Time Monocular Depth Estimation Using Synthetic Data with Domain Adaptation via Image Style Transfer

Atapour-Abarghouei¹,

Breckon

2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

Self Cite

255

173

View full text Add to dashboard Cite

Monocular depth estimation using learning-based approaches has become promising in recent years. However, most monocular depth estimators either need to rely on large quantities of ground truth depth data, which is extremely expensive and difficult to obtain, or predict disparity as an intermediary step using a secondary supervisory signal leading to blurring and other artefacts. Training a depth estimation model using pixel-perfect synthetic data can resolve most of these issues but introduces the problem of domain bias. This is the inability to apply a model trained on synthetic data to real-world scenarios. With advances in image style transfer and its connections with domain adaptation (Maximum Mean Discrepancy), we take advantage of style transfer and adversarial training to predict pixel perfect depth from a single real-world color image based on training over a large corpus of synthetic environment data. Experimental results indicate the efficacy of our approach compared to contemporary state-of-the-art techniques.

show abstract

“…Future work will look to investigate the extension of this approach to the recovery of vehicle and pedestrian interactions for inform human/vehicle activity classification [4,59,62] and also the applicability within the context of mobile platform navigation [63][64][65][66], driver assistance systems [67,68] and for multi-platform, multi-modal wide-area search and surveillance tasks [5,69,70].…”

Section: Discussionmentioning

confidence: 99%