A new pedestrian dataset for supervised learning

Overett, Gary; Petersson, Lars; Brewer, Nathan; Andersson, Lars; Pettersson, Niklas

doi:10.1109/ivs.2008.4621297

Cited by 100 publications

(34 citation statements)

References 16 publications

(12 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We implemented two software simulators in C for both of the schemes and evaluated accuracy of human detection. In this evaluation, NICTA Pedestrian database [15] was used for benchmarking. To generate AdaBoost classifiers, 2,000 images from the database were used for offline machine learning, while other 1,000 images were used as the evaluation data.…”

Section: Implementation Results and Evaluationmentioning

confidence: 99%

FPGA Implementation of Human Detection by HOG Features with AdaBoost

Dohi

Negi

Shibata

et al. 2013

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYWe implement external memory-free deep pipelined FPGA implementation including HOG feature extraction and AdaBoost classification. To construct our design by compact FPGA, we introduce some simplifications of the algorithm and aggressive use of stream oriented architectures. We present comparison results between our simplified fixed-point scheme and an original floating-point scheme in terms of quality of results, and the results suggest the negative impact of the simplified scheme for hardware implementation is limited. We empirically show that, our system is able to detect human from 640 × 480 VGA images at up to 112 FPS on a Xilinx Virtex-5 XC5VLX50 FPGA.

show abstract

Section: Implementation Results and Evaluationmentioning

confidence: 99%

FPGA Implementation of Human Detection by HOG Features with AdaBoost

Dohi

Negi

Shibata

et al. 2013

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…En los próximos tres trabajos la base de datos Daimler pedestrian benchmark (Keller et al, 2011) ha sido empleada para generar los resultados experimentales, así, Min et al (Min et al, 2013) han obtenido una tasa de detección del 73 % a 10 −4 FPPW, Mesmakhosroshahi et al (Mesmakhosroshahi et al, 2014) han alcanzado una tasa de detección del 98,8 % y Zhang et al (Zhang et al, 2015a) han complementado sus experimentos con imágenes capturadas alrededor de la ciudad de Chicago Tetik y Bolat (Tetik and Bolat, 2011) han desarrollado sus experimentos sobre las bases de datos Nicta (Overett et al, 2008) para entrenamiento, y Penn Fudan (Wang et al, 2007) para validación. Esta propuesta alcanza una tasa de detección del 84,4 %.…”

Section: Estado Del Arte En Generación De Rois Y Sdpunclassified

Generación de Regiones con Potencial de Contener Peatones usando Reconstrucción 3D No Densa a partir de Visión Monocular

Zubiaguirre-Bergen

Torres-Torriti

Flores-Calero

2018

Rev. iberoam. autom. inform. ind.

View full text Add to dashboard Cite

ResumenLos accidentes de tráfico son un problema de salud pública a escala mundial, por el alto número de víctimas humanas y los elevados costos económicos y sociales que generan. En este contexto, los peatones se encuentran entre los elementos más importantes y vulnerables de la escena vial que necesitan ser protegidos. Es así que en este trabajo se presenta una innovadora propuesta utilizado la información visual monocular para emular la visión estéreo, y a partir de ello: i) generar regiones de interés (ROIs) con alta posibilidad de contener un peatón, y ii) estimar la trayectoria del vehículo. Los experimentos han sido desarrollados sobre una base de datos de imágenes tomadas en varias calles de la ciudad de Santiago (Región-Metropolitana), Chile. Esta información fue obtenida usando una plataforma experimental en condiciones reales de conducción durante el día. La tasa de detección de ROIs es del 86,6 % para distancias menores a 20 metros, 82,9 % para distancias menores a 30 metros y del 76,2 % para distancias menores a 40 metros.Palabras Clave: Peatones, accidentes, tráfico, visión monocular, visión estéreo, trayectoria, ROIs.Generation of regions of interest with high potential to contain pedestrians using non-dense 3D reconstruction from monocular vision. AbstractTraffic accidents are a global public health problem, due to the high number of human victims and the elevated economic and social costs that generate. In this context, pedestrians are among the most important and vulnerable elements of the road scene that need to be protected. It is thus that, in this work an innovative proposal is presented where the monocular visual information is used to simulate the stereo vision, and from this: i) generate regions of interest (ROIs) with high possibility of containing a pedestrian, and ii) estimate the trajectory of the vehicle. Experiments have been developed into a dataset of images taken in several streets of Santiago (Región Metropolitana), Chile. This database was obtained using an experimental vehicle under real driving conditions during the day. The ROI detection rate is 86,6 % for distances less than 20 meters, 82,9 % for distances less than 30 meters and 76,2 % for distances less than 40 meters. Keywords:Pedestrian, accidents, traffic, monocular vision, stereo vision, trajectory, ROIs. IntroducciónLos accidentes de tráfico causan al año 1,3 millones de muertes y entre 20 y 50 millones de heridos en el mundo, siendo así la primera causa de muerte de jóvenes entre 15 y 29 años de edad. El 91 % de ellos ocurren en países de bajo y mediano ingresos. Por otra parte, niños, peatones, ciclistas y personas de * Autor para correspondencia: mjflores@espe.edu.ec Este artículo ha sido aceptado para su publicación en un futuro número de RIAI. Su contenido es definitivo y únicamente cambiará en la versión final la información relativa al volumen, número y número de páginas.

show abstract

“…These most dominant cells are the cells having the closest HOG vector to the mean HOG vector calculated over the vectors (of the corresponding cell) from a human database. The system was trained using 10, 000 positive and 20, 000 negative image samples from the NICTA database [33]. Figure 3 shows an example of several detected persons in dynamically occluded scenario.…”

Section: Human Detection and Trackingmentioning

confidence: 99%

Boosted human re-identification using Riemannian manifolds

Corvee

Brémond

et al. 2012

Image and Vision Computing

View full text Add to dashboard Cite

This paper presents an appearance-based model to address the human reidentication problem. Human re-identication is an important and still unsolved task in computer vision. In many systems there is a requirement to identify individuals or determine whether a given individual has already appeared somewhere in a network of cameras. The human appearance obtained in one camera is usually dierent from the ones obtained in another camera. In order to re-identify people a human signature should handle dierence in illumination, pose and camera parameters. The paper focuses on a new appearance model based on Mean Riemannian Covariance (MRC) patches extracted from tracks of a particular individual. A new similarity measure using Riemannian manifold theory is also proposed to distinguish sets of patches belonging to a specic individual. We investigate the signicance of the MRC patches based on their reliability extracted during tracking and their discriminative power obtained by a boosting scheme. The methods are evaluated and compared with the state of the art using benchmark video sequences from the ETHZ and the i-LIDS datasets. The re-identication performance is presented using the cumulative matching characteristic (CMC) curve. We demonstrate that the proposed approach outperforms state of the art methods. Finally, the results of our approach are shown on two other more pertinent datasets.

show abstract

A new pedestrian dataset for supervised learning

Cited by 100 publications

References 16 publications

FPGA Implementation of Human Detection by HOG Features with AdaBoost

FPGA Implementation of Human Detection by HOG Features with AdaBoost

Generación de Regiones con Potencial de Contener Peatones usando Reconstrucción 3D No Densa a partir de Visión Monocular

Boosted human re-identification using Riemannian manifolds

Contact Info

Product

Resources

About