Evaluation of stacked autoencoders for pedestrian detection

B. Peralta, L. Parra, L. Caro

Resultado de la investigación: Conference contribution

Resumen

Pedestrian detection has multiple applications as video surveillance, automatic driver-assistance systems in vehicles or visual control of access. This task is challenging due to presence of factors such as poor lighting, occlusion or uncertainty in the environment. Deep learning has reached many state-of-art results in visual recognition, where one popular and simple variant is stacked autoencoders. Nonetheless, it is not clear what is the effect of each stacked autoencoders parameter in pedestrian detection performance. In this work, we propose to revise the feature representation for pedestrian detection considering the use of deep learning using stacked autoencoders with a sensitivity analysis of relevant parameters. Additionally, this paper presents a methodology for feature extraction using stacked autoencoders. The experiments show that this model is capable of creating a meaningful visual descriptor for pedestrian detection, which improves the detection performance in comparison to baseline techniques without an optimal setting of parameters. In presence of occlusion or poor people images, we found diffuse and distorted visual patterns. A future avenue is the learning of the degree of noise for improving the generalization capabilities of the learned features.

Idioma originalEnglish
Título de la publicación alojadaProceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016
EditorialIEEE Computer Society
ISBN (versión digital)9781509033393
DOI
EstadoPublished - 27 ene 2017
Evento35th International Conference of the Chilean Computer Science Society, SCCC 2016 - Valparaiso, Chile
Duración: 10 oct 201614 oct 2016

Conference

Conference35th International Conference of the Chilean Computer Science Society, SCCC 2016
PaísChile
CiudadValparaiso
Período10/10/1614/10/16

Huella dactilar

Sensitivity analysis
Feature extraction
Lighting
Experiments
Deep learning
Uncertainty

ASJC Scopus subject areas

  • Engineering(all)
  • Computer Science(all)

Citar esto

Peralta, B., Parra, L., & Caro, L. (2017). Evaluation of stacked autoencoders for pedestrian detection. En Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016 [7836017] IEEE Computer Society. https://doi.org/10.1109/SCCC.2016.7836017
Peralta, B. ; Parra, L. ; Caro, L. / Evaluation of stacked autoencoders for pedestrian detection. Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016. IEEE Computer Society, 2017.
@inproceedings{340ba97351374369bd90088a253dadf8,
title = "Evaluation of stacked autoencoders for pedestrian detection",
abstract = "Pedestrian detection has multiple applications as video surveillance, automatic driver-assistance systems in vehicles or visual control of access. This task is challenging due to presence of factors such as poor lighting, occlusion or uncertainty in the environment. Deep learning has reached many state-of-art results in visual recognition, where one popular and simple variant is stacked autoencoders. Nonetheless, it is not clear what is the effect of each stacked autoencoders parameter in pedestrian detection performance. In this work, we propose to revise the feature representation for pedestrian detection considering the use of deep learning using stacked autoencoders with a sensitivity analysis of relevant parameters. Additionally, this paper presents a methodology for feature extraction using stacked autoencoders. The experiments show that this model is capable of creating a meaningful visual descriptor for pedestrian detection, which improves the detection performance in comparison to baseline techniques without an optimal setting of parameters. In presence of occlusion or poor people images, we found diffuse and distorted visual patterns. A future avenue is the learning of the degree of noise for improving the generalization capabilities of the learned features.",
keywords = "Autoencoders, Deep Learning, Pedestrian Detection, Stacked Autoencoders",
author = "B. Peralta and L. Parra and L. Caro",
year = "2017",
month = "1",
day = "27",
doi = "10.1109/SCCC.2016.7836017",
language = "English",
booktitle = "Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016",
publisher = "IEEE Computer Society",
address = "United States",

}

Peralta, B, Parra, L & Caro, L 2017, Evaluation of stacked autoencoders for pedestrian detection. En Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016., 7836017, IEEE Computer Society, 35th International Conference of the Chilean Computer Science Society, SCCC 2016, Valparaiso, Chile, 10/10/16. https://doi.org/10.1109/SCCC.2016.7836017

Evaluation of stacked autoencoders for pedestrian detection. / Peralta, B.; Parra, L.; Caro, L.

Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016. IEEE Computer Society, 2017. 7836017.

Resultado de la investigación: Conference contribution

TY - GEN

T1 - Evaluation of stacked autoencoders for pedestrian detection

AU - Peralta, B.

AU - Parra, L.

AU - Caro, L.

PY - 2017/1/27

Y1 - 2017/1/27

N2 - Pedestrian detection has multiple applications as video surveillance, automatic driver-assistance systems in vehicles or visual control of access. This task is challenging due to presence of factors such as poor lighting, occlusion or uncertainty in the environment. Deep learning has reached many state-of-art results in visual recognition, where one popular and simple variant is stacked autoencoders. Nonetheless, it is not clear what is the effect of each stacked autoencoders parameter in pedestrian detection performance. In this work, we propose to revise the feature representation for pedestrian detection considering the use of deep learning using stacked autoencoders with a sensitivity analysis of relevant parameters. Additionally, this paper presents a methodology for feature extraction using stacked autoencoders. The experiments show that this model is capable of creating a meaningful visual descriptor for pedestrian detection, which improves the detection performance in comparison to baseline techniques without an optimal setting of parameters. In presence of occlusion or poor people images, we found diffuse and distorted visual patterns. A future avenue is the learning of the degree of noise for improving the generalization capabilities of the learned features.

AB - Pedestrian detection has multiple applications as video surveillance, automatic driver-assistance systems in vehicles or visual control of access. This task is challenging due to presence of factors such as poor lighting, occlusion or uncertainty in the environment. Deep learning has reached many state-of-art results in visual recognition, where one popular and simple variant is stacked autoencoders. Nonetheless, it is not clear what is the effect of each stacked autoencoders parameter in pedestrian detection performance. In this work, we propose to revise the feature representation for pedestrian detection considering the use of deep learning using stacked autoencoders with a sensitivity analysis of relevant parameters. Additionally, this paper presents a methodology for feature extraction using stacked autoencoders. The experiments show that this model is capable of creating a meaningful visual descriptor for pedestrian detection, which improves the detection performance in comparison to baseline techniques without an optimal setting of parameters. In presence of occlusion or poor people images, we found diffuse and distorted visual patterns. A future avenue is the learning of the degree of noise for improving the generalization capabilities of the learned features.

KW - Autoencoders

KW - Deep Learning

KW - Pedestrian Detection

KW - Stacked Autoencoders

UR - http://www.scopus.com/inward/record.url?scp=85017664060&partnerID=8YFLogxK

U2 - 10.1109/SCCC.2016.7836017

DO - 10.1109/SCCC.2016.7836017

M3 - Conference contribution

BT - Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016

PB - IEEE Computer Society

ER -

Peralta B, Parra L, Caro L. Evaluation of stacked autoencoders for pedestrian detection. En Proceedings of the 2016 35th International Conference of the Chilean Computer Science Society, SCCC 2016. IEEE Computer Society. 2017. 7836017 https://doi.org/10.1109/SCCC.2016.7836017