Sorry, you need to enable JavaScript to visit this website.

Image/Video Processing

FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS


We propose a fixed-map semi-direct visual odometry (FSVO) algorithm for Micro Aerial Vehicles (MAVs). The proposed approach does not need computationally expensive feature extraction and matching techniques for motion estimation at each frame. Instead, we extract and match ORiented Brief (ORB) features between keyframes and assist-frames. We replace the incremental map generation step in traditional algorithms with fixed map generation at keyframe and assist- frame only in our algorithm, resulting in reduced storage memory and higher flexibility for relocalization.

Paper Details

Authors:
Yulan Guo, Zaiping Lin, Wei An
Submitted On:
14 September 2017 - 9:25pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP2017-PosterPresentation.pdf

(14 downloads)

Keywords

Subscribe

[1] Yulan Guo, Zaiping Lin, Wei An, "FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2061. Accessed: Oct. 19, 2017.
@article{2061-17,
url = {http://sigport.org/2061},
author = {Yulan Guo; Zaiping Lin; Wei An },
publisher = {IEEE SigPort},
title = {FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS},
year = {2017} }
TY - EJOUR
T1 - FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS
AU - Yulan Guo; Zaiping Lin; Wei An
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2061
ER -
Yulan Guo, Zaiping Lin, Wei An. (2017). FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS. IEEE SigPort. http://sigport.org/2061
Yulan Guo, Zaiping Lin, Wei An, 2017. FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS. Available at: http://sigport.org/2061.
Yulan Guo, Zaiping Lin, Wei An. (2017). "FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS." Web.
1. Yulan Guo, Zaiping Lin, Wei An. FSVO: SEMI-DIRECT MONOCULAR VISUAL ODOMETRY USING FIXED MAPS [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2061

CSfM: Community-based Structure-from-Motion


Structure-from-Motion approaches could be broadly divided into two classes: incremental and global. While incremental manner is robust to outliers, it suffers from error accumulation and heavy computation load. The global manner has the advantage of simultaneously estimating all camera poses, but it is usually sensitive to epipolar geometry outliers. In this paper, we propose an adaptive community-based SfM (CSfM) method which takes both robustness and efficiency into consideration. First, the epipolar geometry graph is partitioned into separate communities.

Paper Details

Authors:
Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu
Submitted On:
14 September 2017 - 9:20pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP2017_Community-based_SfM_HainanCui

(17 downloads)

Keywords

Subscribe

[1] Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu, "CSfM: Community-based Structure-from-Motion", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2060. Accessed: Oct. 19, 2017.
@article{2060-17,
url = {http://sigport.org/2060},
author = {Hainan Cui; Shuhan Shen; Xiang Gao; Zhanyi Hu },
publisher = {IEEE SigPort},
title = {CSfM: Community-based Structure-from-Motion},
year = {2017} }
TY - EJOUR
T1 - CSfM: Community-based Structure-from-Motion
AU - Hainan Cui; Shuhan Shen; Xiang Gao; Zhanyi Hu
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2060
ER -
Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu. (2017). CSfM: Community-based Structure-from-Motion. IEEE SigPort. http://sigport.org/2060
Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu, 2017. CSfM: Community-based Structure-from-Motion. Available at: http://sigport.org/2060.
Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu. (2017). "CSfM: Community-based Structure-from-Motion." Web.
1. Hainan Cui, Shuhan Shen, Xiang Gao, Zhanyi Hu. CSfM: Community-based Structure-from-Motion [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2060

Semantic Background Subtraction


We introduce the notion of semantic background subtraction, a novel framework for motion detection in video sequences. The key innovation consists to leverage object-level semantics to address the variety of challenging scenarios for background subtraction. Our framework combines the information of a semantic segmentation algorithm, expressed by a probability for each pixel, with the output of any background subtraction algorithm to reduce false positive detections produced by illumination changes, dynamic backgrounds, strong shadows, and ghosts.

Paper Details

Authors:
Sébastien Piérard, Marc Van Droogenbroeck
Submitted On:
14 September 2017 - 8:44pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Braham_ICIP2017_poster.pdf

(14 downloads)

Keywords

Subscribe

[1] Sébastien Piérard, Marc Van Droogenbroeck, "Semantic Background Subtraction", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2059. Accessed: Oct. 19, 2017.
@article{2059-17,
url = {http://sigport.org/2059},
author = {Sébastien Piérard; Marc Van Droogenbroeck },
publisher = {IEEE SigPort},
title = {Semantic Background Subtraction},
year = {2017} }
TY - EJOUR
T1 - Semantic Background Subtraction
AU - Sébastien Piérard; Marc Van Droogenbroeck
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2059
ER -
Sébastien Piérard, Marc Van Droogenbroeck. (2017). Semantic Background Subtraction. IEEE SigPort. http://sigport.org/2059
Sébastien Piérard, Marc Van Droogenbroeck, 2017. Semantic Background Subtraction. Available at: http://sigport.org/2059.
Sébastien Piérard, Marc Van Droogenbroeck. (2017). "Semantic Background Subtraction." Web.
1. Sébastien Piérard, Marc Van Droogenbroeck. Semantic Background Subtraction [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2059

COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”


We investigate video classification via a 3D deep convolutional neural network (CNN) that directly ingests compressed bitstream information. This idea is based on the observation that video macroblock (MB) motion vectors (that are very compact and directly available from the compressed bitstream) are inherently capturing local spatiotemporal changes in each video scene.

Paper Details

Authors:
Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos
Submitted On:
14 September 2017 - 8:03pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Compressed_domain_video_classification.pdf

(31 downloads)

Keywords

Subscribe

[1] Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos, "COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2057. Accessed: Oct. 19, 2017.
@article{2057-17,
url = {http://sigport.org/2057},
author = {Aaron Chadha; Alhabib Abbas; Yiannis Andreopoulos },
publisher = {IEEE SigPort},
title = {COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”},
year = {2017} }
TY - EJOUR
T1 - COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”
AU - Aaron Chadha; Alhabib Abbas; Yiannis Andreopoulos
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2057
ER -
Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos. (2017). COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”. IEEE SigPort. http://sigport.org/2057
Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos, 2017. COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”. Available at: http://sigport.org/2057.
Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos. (2017). "COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX”." Web.
1. Aaron Chadha, Alhabib Abbas, Yiannis Andreopoulos. COMPRESSED-DOMAIN VIDEO CLASSIFICATION WITH DEEP NEURAL NETWORKS: “THERE’S WAY TOO MUCH INFORMATION TO DECODE THE MATRIX” [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2057

Convolutional Neural Networks and Training Strategies for Skin Detection

Paper Details

Authors:
Submitted On:
14 September 2017 - 8:01pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

[Poster] ICIP_Yoonsik Kim_20170915.pdf

(15 downloads)

Keywords

Subscribe

[1] , "Convolutional Neural Networks and Training Strategies for Skin Detection", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2056. Accessed: Oct. 19, 2017.
@article{2056-17,
url = {http://sigport.org/2056},
author = { },
publisher = {IEEE SigPort},
title = {Convolutional Neural Networks and Training Strategies for Skin Detection},
year = {2017} }
TY - EJOUR
T1 - Convolutional Neural Networks and Training Strategies for Skin Detection
AU -
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2056
ER -
. (2017). Convolutional Neural Networks and Training Strategies for Skin Detection. IEEE SigPort. http://sigport.org/2056
, 2017. Convolutional Neural Networks and Training Strategies for Skin Detection. Available at: http://sigport.org/2056.
. (2017). "Convolutional Neural Networks and Training Strategies for Skin Detection." Web.
1. . Convolutional Neural Networks and Training Strategies for Skin Detection [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2056

MOTION-CONSISTENT VIDEO INPAINTING


This project aims to propose a fast and automatic inpainting technique for high-definition videos which works under many challenging conditions such as a moving camera, a dynamic background or a long occlusion. Our algorithm does not limit to objects removal type but extends to simultaneous background and foreground reconstruction even when the moving objects are occluded for a long period. Built upon Newson et al [1] which optimizes a global patch-based function, our method holds a significant improvement by the introduction of the optical flow term.

final.pptx

File final.pptx (12 downloads)

Paper Details

Authors:
Andres Almansa, Yann Gousseau, Simon Masnou
Submitted On:
14 September 2017 - 8:06pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

final.pptx

(12 downloads)

Keywords

Subscribe

[1] Andres Almansa, Yann Gousseau, Simon Masnou, "MOTION-CONSISTENT VIDEO INPAINTING", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2054. Accessed: Oct. 19, 2017.
@article{2054-17,
url = {http://sigport.org/2054},
author = {Andres Almansa; Yann Gousseau; Simon Masnou },
publisher = {IEEE SigPort},
title = {MOTION-CONSISTENT VIDEO INPAINTING},
year = {2017} }
TY - EJOUR
T1 - MOTION-CONSISTENT VIDEO INPAINTING
AU - Andres Almansa; Yann Gousseau; Simon Masnou
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2054
ER -
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). MOTION-CONSISTENT VIDEO INPAINTING. IEEE SigPort. http://sigport.org/2054
Andres Almansa, Yann Gousseau, Simon Masnou, 2017. MOTION-CONSISTENT VIDEO INPAINTING. Available at: http://sigport.org/2054.
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). "MOTION-CONSISTENT VIDEO INPAINTING." Web.
1. Andres Almansa, Yann Gousseau, Simon Masnou. MOTION-CONSISTENT VIDEO INPAINTING [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2054

MOTION-CONSISTENT VIDEO INPAINTING


This demonstration aims to show some resulting videos for our method presented in ICIP. It is a fast and automatic inpainting technique for high-definition videos which works under many challenging conditions such as a moving camera, a dynamic background or a long-lasting occlusion. By incorporating optical flow in a global patch-based algorithm, our method provide improvements compared to the state-of-the-art, especially in motion preservation.

Paper Details

Authors:
Andres Almansa, Yann Gousseau, Simon Masnou
Submitted On:
14 September 2017 - 5:32pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster_ThucTrinhLE_videoinpainting_ICIP.pdf

(10 downloads)

Keywords

Subscribe

[1] Andres Almansa, Yann Gousseau, Simon Masnou, "MOTION-CONSISTENT VIDEO INPAINTING", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2050. Accessed: Oct. 19, 2017.
@article{2050-17,
url = {http://sigport.org/2050},
author = {Andres Almansa; Yann Gousseau; Simon Masnou },
publisher = {IEEE SigPort},
title = {MOTION-CONSISTENT VIDEO INPAINTING},
year = {2017} }
TY - EJOUR
T1 - MOTION-CONSISTENT VIDEO INPAINTING
AU - Andres Almansa; Yann Gousseau; Simon Masnou
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2050
ER -
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). MOTION-CONSISTENT VIDEO INPAINTING. IEEE SigPort. http://sigport.org/2050
Andres Almansa, Yann Gousseau, Simon Masnou, 2017. MOTION-CONSISTENT VIDEO INPAINTING. Available at: http://sigport.org/2050.
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). "MOTION-CONSISTENT VIDEO INPAINTING." Web.
1. Andres Almansa, Yann Gousseau, Simon Masnou. MOTION-CONSISTENT VIDEO INPAINTING [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2050

MOTION-CONSISTENT VIDEO INPAINTING


This demonstration aims to show some resulting videos for our method presented in ICIP. It is a fast and automatic inpainting technique for high-definition videos which works under many challenging conditions such as a moving camera, a dynamic background or a long-lasting occlusion. By incorporating optical flow in a global patch-based algorithm, our method provide improvements compared to the state-of-the-art, especially in motion preservation.

Paper Details

Authors:
Andres Almansa, Yann Gousseau, Simon Masnou
Submitted On:
14 September 2017 - 5:32pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Poster_ThucTrinhLE_videoinpainting_ICIP.pdf

(10 downloads)

Keywords

Subscribe

[1] Andres Almansa, Yann Gousseau, Simon Masnou, "MOTION-CONSISTENT VIDEO INPAINTING", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2049. Accessed: Oct. 19, 2017.
@article{2049-17,
url = {http://sigport.org/2049},
author = {Andres Almansa; Yann Gousseau; Simon Masnou },
publisher = {IEEE SigPort},
title = {MOTION-CONSISTENT VIDEO INPAINTING},
year = {2017} }
TY - EJOUR
T1 - MOTION-CONSISTENT VIDEO INPAINTING
AU - Andres Almansa; Yann Gousseau; Simon Masnou
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2049
ER -
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). MOTION-CONSISTENT VIDEO INPAINTING. IEEE SigPort. http://sigport.org/2049
Andres Almansa, Yann Gousseau, Simon Masnou, 2017. MOTION-CONSISTENT VIDEO INPAINTING. Available at: http://sigport.org/2049.
Andres Almansa, Yann Gousseau, Simon Masnou. (2017). "MOTION-CONSISTENT VIDEO INPAINTING." Web.
1. Andres Almansa, Yann Gousseau, Simon Masnou. MOTION-CONSISTENT VIDEO INPAINTING [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2049

ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR

Paper Details

Authors:
Lindsey Ravens, Dah-Jye Lee, Alok Desai
Submitted On:
14 September 2017 - 5:12pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

SYBA ICIP Presentation.pdf

(16 downloads)

Keywords

Subscribe

[1] Lindsey Ravens, Dah-Jye Lee, Alok Desai, "ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2047. Accessed: Oct. 19, 2017.
@article{2047-17,
url = {http://sigport.org/2047},
author = {Lindsey Ravens; Dah-Jye Lee; Alok Desai },
publisher = {IEEE SigPort},
title = {ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR},
year = {2017} }
TY - EJOUR
T1 - ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR
AU - Lindsey Ravens; Dah-Jye Lee; Alok Desai
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2047
ER -
Lindsey Ravens, Dah-Jye Lee, Alok Desai. (2017). ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR. IEEE SigPort. http://sigport.org/2047
Lindsey Ravens, Dah-Jye Lee, Alok Desai, 2017. ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR. Available at: http://sigport.org/2047.
Lindsey Ravens, Dah-Jye Lee, Alok Desai. (2017). "ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR." Web.
1. Lindsey Ravens, Dah-Jye Lee, Alok Desai. ROBUST SYNTHETIC BASIS FEATURE DESCRIPTOR [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2047

Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet

Paper Details

Authors:
Submitted On:
14 September 2017 - 12:11pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

icip2017-poster-zhipeng.pdf

(14 downloads)

Keywords

Additional Categories

Subscribe

[1] , "Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2039. Accessed: Oct. 19, 2017.
@article{2039-17,
url = {http://sigport.org/2039},
author = { },
publisher = {IEEE SigPort},
title = {Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet},
year = {2017} }
TY - EJOUR
T1 - Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet
AU -
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2039
ER -
. (2017). Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet. IEEE SigPort. http://sigport.org/2039
, 2017. Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet. Available at: http://sigport.org/2039.
. (2017). "Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet." Web.
1. . Semantic Segmentation with Multi-path Refinement and Pyramid Pooling Dilated-Resnet [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2039

Pages