A Third Eye for Performance Evaluation in Stereo Sequence Analysis

Morales, Sandino; Klette, Reinhard

doi:10.1007/978-3-642-03767-2_131

Cited by 33 publications

(16 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another option is to exploit a prior over the data-set, such as the presence of freespace in front of a manually driven vehicle [5] to detect wrongly reconstructed points (false stereo correspondences) over an extended period of time. Finally, a virtual view synthesized from the reconstructed environment geometry can be compared with the actual data recorded by a third, suitably positioned physical camera [2], [6]. In this evaluation the algorithms have been tested using all the mentioned approaches, in order to get a better understanding of their behavior in real-world applications.…”

Section: Methodsmentioning

confidence: 99%

“…As an alternative, the use of a third camera [6] allows to directly compare a reconstructed view with the actual images without any manual intervention. The computed disparity map is used to transform image pixels taken from the reference camera into control camera coordinates, thus effectively creating a virtual image that can be compared with that recorded by the control camera to produce a cross correlation map.…”

Section: Normalized Cross Correlationmentioning

confidence: 99%

“…The computed disparity map is used to transform image pixels taken from the reference camera into control camera coordinates, thus effectively creating a virtual image that can be compared with that recorded by the control camera to produce a cross correlation map. The chosen metric is the Normalized Cross Correlation (NCC), computed as described in [6]. It is worth noticing that in [2] it is suggested a configuration with the reference and match camera lying 30 cm apart, and the control camera at 50 cm from the reference camera; however, the recording platform used in this work uses much shorter distances (24 and 12 cm respectively) since it has been equipped with a pre-calibrated trinocular camera.…”

Section: Normalized Cross Correlationmentioning

confidence: 99%

See 2 more Smart Citations

Performance analysis of stereo reconstruction algorithms

Morales

Camellini

Felisa

et al. 2013

16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013)

View full text Add to dashboard Cite

Abstract-Environment mapping is one of the most critical tasks in the development of driving assistance systems and stereo vision has been widely used to accomplish it. However, there are very few datasets that allow assessing the performance of a specific method in a real world application. Most datasets have been created in controlled conditions, thus neglecting scenarios that are impossible to reproduce in a laboratory. In this paper, we present the results of the evaluation of three different dense reconstruction algorithm implementations using a number of well-known strategies that represent different tradeoffs in terms of cost, set up time and accuracy. In our tests, we evaluated two variants of the Semi-Global Matching algorithm, and the Efficient Large-Scale Stereo Matching method, as well as different combinations of additional filters in order to assess their influence on the final behavior of the algorithms.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Normalized Cross Correlationmentioning

confidence: 99%

Section: Normalized Cross Correlationmentioning

confidence: 99%

See 1 more Smart Citation

Performance analysis of stereo reconstruction algorithms

Morales

Camellini

Felisa

et al. 2013

16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013)

View full text Add to dashboard Cite

show abstract

“…The third-eye method [11] requires that three calibrated cameras record timesynchronized the same scene. In case of Set 9 on EISATS [4], the cameras are placed on a bar on the left, center, and right position behind the windscreen of the ego-vehicle (i.e.…”

Section: Fundamentalsmentioning

confidence: 99%

“…The third-eye method of [11] provides stereo analysis performance evaluation for long or continuously recorded stereo sequences. For a current application of this method, see [12].…”

Section: Introductionmentioning

confidence: 99%

Third-Eye Stereo Analysis Evaluation Enhanced by Data Measures

Suaste

Caudillo

Shin

et al. 2013

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. Third-eye stereo analysis evaluation compares a virtual image, derived from results obtained by binocular stereo analysis, with a recorded image at the same pose. This technique is applied for evaluating stereo matchers on long (or continuous) stereo input sequences where no ground truth is available. The paper provides a critical and constructive discussion of this method. The paper also introduces data measures on input video sequences as an additional tool for analyzing issues of stereo matchers occurring for particular scenarios. The paper also reports on extensive experiments using two top-rated stereo matchers.

show abstract

Vision‐Based Driver Assistance

Klette

2015

Wiley Encyclopedia of Electrical and Electronics Engineering

Self Cite

View full text Add to dashboard Cite

Vision‐based driver assistance systems (VB‐DAS) still assume a driver in the ego‐vehicle (i.e., that vehicle where the system operates in); autonomous driving can benefit from solutions designed for VB‐DAS. The article reviews research subjects for VB‐DAS and provides an extensive bibliography. It starts with defining basic concepts and notations in the field, also highlighting the importance of performance evaluation for VB‐DAS solutions. The article then reviews briefly safety and comfort functionalities (avoidance of blind spots, night vision, and virtual windshield), before discussing in detail solutions for basic and midlevel environment perception. This involves distance and motion computation, ego‐motion understanding, obstacle detection, tasks of object tracking (vehicles or pedestrians), and the detection and recognition of infrastructure key elements (road, lane, traffic signs, and free space). Representative VB‐DAS examples (e.g., driver monitoring and driver‐environment understanding, speed adaptation, queuing, parking, blind spot supervision, lane departure warning, wrong lane detection, intelligent headlamp control, or intercar communication) are also described. Outlook toward autonomous driving and road‐environment modeling conclude the article.

show abstract

A Third Eye for Performance Evaluation in Stereo Sequence Analysis

Cited by 33 publications

References 9 publications

Performance analysis of stereo reconstruction algorithms

Performance analysis of stereo reconstruction algorithms

Third-Eye Stereo Analysis Evaluation Enhanced by Data Measures

Vision‐Based Driver Assistance

Contact Info

Product

Resources

About