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

Neural network learning (MLR-NNLR)

On Regression Losses for Depth Estimation


Depth estimation from a single monocular image has reached great performances thanks to recent works based on deep networks. However, as various choices of losses, architectures and experimental conditions are proposed in the literature, it is difficult to establish their respective influence on the performances. In this paper we propose an in-depth study of various losses and experimental conditions for depth regression, on \nyu dataset. From this study we propose a new network for depth estimation combining an encoder-decoder architecture with an adversarial loss.

Paper Details

Authors:
Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat
Submitted On:
9 October 2018 - 7:18am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

icip2018_id1083.pdf

(105)

Subscribe

[1] Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat, "On Regression Losses for Depth Estimation", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3656. Accessed: Jul. 23, 2019.
@article{3656-18,
url = {http://sigport.org/3656},
author = {Marcela Carvalho; Bertrand Le Saux; Pauline Trouvé-Peloux; Andrés Almansa; Frédéric Champagnat },
publisher = {IEEE SigPort},
title = {On Regression Losses for Depth Estimation},
year = {2018} }
TY - EJOUR
T1 - On Regression Losses for Depth Estimation
AU - Marcela Carvalho; Bertrand Le Saux; Pauline Trouvé-Peloux; Andrés Almansa; Frédéric Champagnat
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3656
ER -
Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat. (2018). On Regression Losses for Depth Estimation. IEEE SigPort. http://sigport.org/3656
Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat, 2018. On Regression Losses for Depth Estimation. Available at: http://sigport.org/3656.
Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat. (2018). "On Regression Losses for Depth Estimation." Web.
1. Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Frédéric Champagnat. On Regression Losses for Depth Estimation [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3656

Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning


• To automatically segment optic disk (OD) and cup regions in fundus images to derive clinical parameters, such as, cup-to-disk diameter ratio (CDR), to assist glaucoma diagnosis. An eye fundus camera is non-invasive and low-cost,
enabling the screening of a large number of patients quickly.

• Discuss various strategies on how to leverage multiple doctor annotations and prioritize pixels belonging to different regions during network optimization.

• Evaluate proposed approaches on the Drishti-GS dataset.

Paper Details

Authors:
Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale
Submitted On:
8 October 2018 - 7:18pm
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP_2018_Poster.pdf

(91)

Subscribe

[1] Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale, "Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3650. Accessed: Jul. 23, 2019.
@article{3650-18,
url = {http://sigport.org/3650},
author = {Venkata Gopal Edupuganti; Akshay Chawla and Amit Kale },
publisher = {IEEE SigPort},
title = {Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning},
year = {2018} }
TY - EJOUR
T1 - Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning
AU - Venkata Gopal Edupuganti; Akshay Chawla and Amit Kale
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3650
ER -
Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale. (2018). Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning. IEEE SigPort. http://sigport.org/3650
Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale, 2018. Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning. Available at: http://sigport.org/3650.
Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale. (2018). "Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning." Web.
1. Venkata Gopal Edupuganti, Akshay Chawla and Amit Kale. Automatic Optic Disk and Cup Segmentation of Fundus Images Using Deep Learning [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3650

DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION

Paper Details

Authors:
Jianrui Cai
Submitted On:
6 October 2018 - 2:40am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Presentation_ICIP_CJR.pdf

(77)

Subscribe

[1] Jianrui Cai, "DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3564. Accessed: Jul. 23, 2019.
@article{3564-18,
url = {http://sigport.org/3564},
author = {Jianrui Cai },
publisher = {IEEE SigPort},
title = {DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION},
year = {2018} }
TY - EJOUR
T1 - DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION
AU - Jianrui Cai
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3564
ER -
Jianrui Cai. (2018). DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION. IEEE SigPort. http://sigport.org/3564
Jianrui Cai, 2018. DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION. Available at: http://sigport.org/3564.
Jianrui Cai. (2018). "DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION." Web.
1. Jianrui Cai. DEEP IMAGE COMPRESSION WITH ITERATIVE NON-UNIFORM QUANTIZATION [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3564

Fully Convolutional Siamese Networks for Change Detection


This paper presents three fully convolutional neural network architectures which perform change detection using a pair of coregistered images. Most notably, we propose two Siamese extensions of fully convolutional networks which use heuristics about the current problem to achieve the best results in our tests on two open change detection datasets, using both RGB and multispectral images. We show that our system is able to learn from scratch using annotated change detection images.

Paper Details

Authors:
Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch
Submitted On:
5 October 2018 - 5:03am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

_ICIP2018__Poster.pdf

(77)

Subscribe

[1] Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, "Fully Convolutional Siamese Networks for Change Detection", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3524. Accessed: Jul. 23, 2019.
@article{3524-18,
url = {http://sigport.org/3524},
author = {Rodrigo Caye Daudt; Bertrand Le Saux; Alexandre Boulch },
publisher = {IEEE SigPort},
title = {Fully Convolutional Siamese Networks for Change Detection},
year = {2018} }
TY - EJOUR
T1 - Fully Convolutional Siamese Networks for Change Detection
AU - Rodrigo Caye Daudt; Bertrand Le Saux; Alexandre Boulch
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3524
ER -
Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch. (2018). Fully Convolutional Siamese Networks for Change Detection. IEEE SigPort. http://sigport.org/3524
Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, 2018. Fully Convolutional Siamese Networks for Change Detection. Available at: http://sigport.org/3524.
Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch. (2018). "Fully Convolutional Siamese Networks for Change Detection." Web.
1. Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch. Fully Convolutional Siamese Networks for Change Detection [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3524

Learning Illuminant Estimation from Object Recognition


In this paper we present a deep learning method to estimate the illuminant of an image. Our model is not trained with illuminant annotations, but with the objective of improving performance on an auxiliary task such as object recognition. To the best of our knowledge, this is the first example of a deep learning architecture for illuminant estimation that is trained without ground truth illuminants. We evaluate our solution on standard datasets for color constancy, and compare it with state of the art methods.

Paper Details

Authors:
Marco Buzzelli, Joost van de Weijer, Raimondo Schettini
Submitted On:
5 October 2018 - 3:02am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Learning Illuminant Estimation from Object Recognition

(55)

Subscribe

[1] Marco Buzzelli, Joost van de Weijer, Raimondo Schettini, "Learning Illuminant Estimation from Object Recognition", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3507. Accessed: Jul. 23, 2019.
@article{3507-18,
url = {http://sigport.org/3507},
author = {Marco Buzzelli; Joost van de Weijer; Raimondo Schettini },
publisher = {IEEE SigPort},
title = {Learning Illuminant Estimation from Object Recognition},
year = {2018} }
TY - EJOUR
T1 - Learning Illuminant Estimation from Object Recognition
AU - Marco Buzzelli; Joost van de Weijer; Raimondo Schettini
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3507
ER -
Marco Buzzelli, Joost van de Weijer, Raimondo Schettini. (2018). Learning Illuminant Estimation from Object Recognition. IEEE SigPort. http://sigport.org/3507
Marco Buzzelli, Joost van de Weijer, Raimondo Schettini, 2018. Learning Illuminant Estimation from Object Recognition. Available at: http://sigport.org/3507.
Marco Buzzelli, Joost van de Weijer, Raimondo Schettini. (2018). "Learning Illuminant Estimation from Object Recognition." Web.
1. Marco Buzzelli, Joost van de Weijer, Raimondo Schettini. Learning Illuminant Estimation from Object Recognition [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3507

S3D: Stacking Segmental P3D for Action Quality Assessment


Action quality assessment is crucial in areas of sports, surgery and assembly line where action skills can be evaluated. In this paper, we propose the Segment-based P3D-fused network S3D built-upon ED-TCN and push the performance on the UNLV-Dive dataset by a significant margin. We verify that segment-aware training performs better than full-video training which turns out to focus on the water spray. We show that temporal segmentation can be embedded with few efforts.

Paper Details

Authors:
Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran
Submitted On:
5 October 2018 - 2:08am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

AI Referee: Score Olympic Games

(118)

Subscribe

[1] Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran, "S3D: Stacking Segmental P3D for Action Quality Assessment", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3501. Accessed: Jul. 23, 2019.
@article{3501-18,
url = {http://sigport.org/3501},
author = {Ye Tian; Austin Reiter; Gregory D. Hager; Trac D. Tran },
publisher = {IEEE SigPort},
title = {S3D: Stacking Segmental P3D for Action Quality Assessment},
year = {2018} }
TY - EJOUR
T1 - S3D: Stacking Segmental P3D for Action Quality Assessment
AU - Ye Tian; Austin Reiter; Gregory D. Hager; Trac D. Tran
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3501
ER -
Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran. (2018). S3D: Stacking Segmental P3D for Action Quality Assessment. IEEE SigPort. http://sigport.org/3501
Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran, 2018. S3D: Stacking Segmental P3D for Action Quality Assessment. Available at: http://sigport.org/3501.
Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran. (2018). "S3D: Stacking Segmental P3D for Action Quality Assessment." Web.
1. Ye Tian, Austin Reiter, Gregory D. Hager, Trac D. Tran. S3D: Stacking Segmental P3D for Action Quality Assessment [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3501

Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection


One of the most important factors in training object recognition networks using convolutional neural networks (CNN) is the provision of annotated data accompanying human judgment. Particularly, in object detection or semantic segmentation, the annotation process requires considerable human effort. In this paper, we propose a semi-supervised learning (SSL)-based training methodology for object detection, which makes use of automatic labeling of un-annotated data by applying a network previously trained from an annotated dataset.

Paper Details

Authors:
Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon
Submitted On:
5 October 2018 - 2:00am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection.pdf

(14)

Subscribe

[1] Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon, "Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3499. Accessed: Jul. 23, 2019.
@article{3499-18,
url = {http://sigport.org/3499},
author = {Min-Kook Choi; Jaehyeong Park; Jihun Jung; Heechul Jung; Jin-Hee Lee; Woong Jae Won; Woo Young Jung; Jincheol Kim; Soon Kwon },
publisher = {IEEE SigPort},
title = {Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection},
year = {2018} }
TY - EJOUR
T1 - Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection
AU - Min-Kook Choi; Jaehyeong Park; Jihun Jung; Heechul Jung; Jin-Hee Lee; Woong Jae Won; Woo Young Jung; Jincheol Kim; Soon Kwon
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3499
ER -
Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon. (2018). Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection. IEEE SigPort. http://sigport.org/3499
Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon, 2018. Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection. Available at: http://sigport.org/3499.
Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon. (2018). "Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection." Web.
1. Min-Kook Choi, Jaehyeong Park, Jihun Jung, Heechul Jung, Jin-Hee Lee, Woong Jae Won, Woo Young Jung, Jincheol Kim, Soon Kwon. Co-occurrence Matrix Analysis Based Semi-Supervised Training for Object Detection [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3499

Semantically Interpretable and Controllable Filter Sets


In this paper, we generate and control semantically interpretable filters that are directly learned from natural images in an unsupervised fashion. Each semantic filter learns a visually interpretable local structure in conjunction with other filters. The significance of learning these interpretable filter sets is demonstrated on two contrasting applications. The first application is image recognition under progressive decolorization, in which recognition algorithms should be color-insensitive to achieve a robust performance.

Paper Details

Authors:
Gukyeong Kwon
Submitted On:
5 October 2018 - 12:19am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

SemAE_Poster

(63)

Subscribe

[1] Gukyeong Kwon, "Semantically Interpretable and Controllable Filter Sets", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3490. Accessed: Jul. 23, 2019.
@article{3490-18,
url = {http://sigport.org/3490},
author = {Gukyeong Kwon },
publisher = {IEEE SigPort},
title = {Semantically Interpretable and Controllable Filter Sets},
year = {2018} }
TY - EJOUR
T1 - Semantically Interpretable and Controllable Filter Sets
AU - Gukyeong Kwon
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3490
ER -
Gukyeong Kwon. (2018). Semantically Interpretable and Controllable Filter Sets. IEEE SigPort. http://sigport.org/3490
Gukyeong Kwon, 2018. Semantically Interpretable and Controllable Filter Sets. Available at: http://sigport.org/3490.
Gukyeong Kwon. (2018). "Semantically Interpretable and Controllable Filter Sets." Web.
1. Gukyeong Kwon. Semantically Interpretable and Controllable Filter Sets [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3490

TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching

Paper Details

Authors:
Sovann EN, Alexis LECHERVY, Frédéric JURIE
Submitted On:
4 October 2018 - 9:18am
Short Link:
Type:
Event:
Presenter's Name:
Paper Code:
Document Year:
Cite

Document Files

ICIP_18_poster.pdf

(67)

Subscribe

[1] Sovann EN, Alexis LECHERVY, Frédéric JURIE, "TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3395. Accessed: Jul. 23, 2019.
@article{3395-18,
url = {http://sigport.org/3395},
author = {Sovann EN; Alexis LECHERVY; Frédéric JURIE },
publisher = {IEEE SigPort},
title = {TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching},
year = {2018} }
TY - EJOUR
T1 - TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching
AU - Sovann EN; Alexis LECHERVY; Frédéric JURIE
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3395
ER -
Sovann EN, Alexis LECHERVY, Frédéric JURIE. (2018). TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching. IEEE SigPort. http://sigport.org/3395
Sovann EN, Alexis LECHERVY, Frédéric JURIE, 2018. TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching. Available at: http://sigport.org/3395.
Sovann EN, Alexis LECHERVY, Frédéric JURIE. (2018). "TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching." Web.
1. Sovann EN, Alexis LECHERVY, Frédéric JURIE. TS-Net: Combining Modality Specific and Common Features for Multimodal Patch Matching [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3395

WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS


The task of rain detection, or wet-dry classification
using measurements from commercial microwave links (CMLs)
is a subject that been studied in depth. However, these studies
are based on direct measurement of the signal level, which
is known to be attenuated by rain. In this paper we present,
for the first time an empirical study on rain classification using
records of transmissions errors in the CMLs. Based on a dataset
of measurements taken from operational cellular backhaul
networks and meteorological measurements, and using long

Paper Details

Authors:
Hagit Messer
Submitted On:
2 July 2018 - 11:34am
Short Link:
Type:
Event:
Presenter's Name:
Document Year:
Cite

Document Files

poster-sam2018.pdf

(106)

Subscribe

[1] Hagit Messer, "WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3349. Accessed: Jul. 23, 2019.
@article{3349-18,
url = {http://sigport.org/3349},
author = {Hagit Messer },
publisher = {IEEE SigPort},
title = {WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS},
year = {2018} }
TY - EJOUR
T1 - WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS
AU - Hagit Messer
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3349
ER -
Hagit Messer. (2018). WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS. IEEE SigPort. http://sigport.org/3349
Hagit Messer, 2018. WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS. Available at: http://sigport.org/3349.
Hagit Messer. (2018). "WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS." Web.
1. Hagit Messer. WET-DRY CLASSIFICATION USING LSTM AND COMMERCIALMICROWAVE LINKS [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3349

Pages