A time delay neural network algorithm for estimating image-pattern shape and motion

Wöhler, Christian; Anlauf, Joachim K.

doi:10.1016/s0262-8856(98)00108-5

Cited by 33 publications

(37 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, local binary pattern (LBP) features [39] have also been employed in pedestrian classification [53]. The particular structure of local texture features has been optimized in terms of local receptive field (LRF) features [11], [19], [40], [55], which adapt to the underlying data during training. Other texture-based features are codebook patches, extracted around interest points in the image [1], [28], [45] and linked via geometric relations.…”

Section: Previous Workmentioning

confidence: 99%

See 1 more Smart Citation

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

Enzweiler

Gavrila

2011

IEEE Trans. on Image Process.

127

View full text Add to dashboard Cite

Abstract-Notwithstanding many years of progress, pedestrian recognition is still a difficult but important problem. We present a novel multilevel Mixture-of-Experts approach to combine information from multiple features and cues with the objective of improved pedestrian classification. On pose-level, shape cues based on Chamfer shape matching provide sample-dependent priors for a certain pedestrian view. On modality-level, we represent each data sample in terms of image intensity, (dense) depth, and (dense) flow. On feature-level, we consider histograms of oriented gradients (HOG) and local binary patterns (LBP). Multilayer perceptrons (MLP) and linear support vector machines (linSVM) are used as expert classifiers.Experiments are performed on a unique real-world multimodality dataset captured from a moving vehicle in urban traffic. This dataset has been made public for research purposes. Our results show a significant performance boost of up to a factor of 42 in reduction of false positives at constant detection rates of our approach compared to a baseline intensity-only HOG/linSVM approach.

show abstract

Section: Previous Workmentioning

confidence: 99%

“…Some of the presented spatial filters have been extended to the spatio-temporal domain by means of intensity differences over time [50], [55] or optical flow [5].…”

Section: Previous Workmentioning

confidence: 99%

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

Enzweiler

Gavrila

2011

IEEE Trans. on Image Process.

127

View full text Add to dashboard Cite

show abstract

“…Nonadaptive Haar wavelet features have been popularized by [35] and adapted by many others [28], [42], with manual [28], [35] and automatic feature selection [42]. Adaptive feature sets were proposed, e.g., local receptive fields (LRFs) [45], where the spatial structure is able to adapt to the data. Another class involves code-book patches that are extracted around salient points in the image, e.g., [25].…”

Section: Previous Workmentioning

confidence: 99%

“…Other popular classifiers include neural networks [15], [45] and AdaBoost cascades [27], [40], [42], [46], [47], [49], [51]. Some approaches additionally apply a component-based representation of pedestrians as an ensemble of body parts [8], [27], [28], [47].…”

Section: Previous Workmentioning

confidence: 99%

The Benefits of Dense Stereo for Pedestrian Detection

Keller

Enzweiler

Rohrbach

et al. 2011

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

Abstract-This paper presents a novel pedestrian detection system for intelligent vehicles. We propose the use of dense stereo for both the generation of regions of interest and pedestrian classification. Dense stereo allows the dynamic estimation of camera parameters and the road profile, which, in turn, provides strong scene constraints on possible pedestrian locations. For classification, we extract spatial features (gradient orientation histograms) directly from dense depth and intensity images. Both modalities are represented in terms of individual feature spaces, in which discriminative classifiers (linear support vector machines) are learned. We refrain from the construction of a joint feature space but instead employ a fusion of depth and intensity on the classifier level. Our experiments involve challenging image data captured in complex urban environments (i.e., undulating roads and speed bumps). Our results show a performance improvement by up to a factor of 7.5 at the classification level and up to a factor of 5 at the tracking level (reduction in false alarms at constant detection rates) over a system with static scene constraints and intensity-only classification.

show abstract

“…The concept of an This paper is the extended version from paper titled "Detection of Visual Bearing Defect Using Integrated Artificial Neural Network" that has been published in Proceeding of ICACSIS 2011. artificial neural network architecture with time delay also adopt similar concept. This architecture can be used in finding boundaries, direction and speed of motion of moving images [3].…”

Section: Introductionmentioning

confidence: 99%

Untitled

Herdianta¹,

Nasution²

2012

Jurnal Ilmu Komputer dan Informasi

View full text Add to dashboard Cite

Defects on a bearing is usually determined by observing its vibration characteristics. This method unfortunately can not detect the visual defects on the inner and outer ring bearing surface. A pattern recognition is implemented in this paper to solve the problem. A backpropagation neural network architecture is used to recognize the visual defect pattern. This architecture is integrated in a digital image processing chain. Recognition rate of good bearing is obtained at 92.93 %, meanwhile for defected bearing is obtained at 75 % respectively. This rate shows integrated artificial neural network with digital image processing can be implemented to detect the presence of visual bearing defect. Keywords: backpropagation, bearing, visual defect AbstrakCacat pada bearing biasanya ditentukan dengan mengamati karakteristik getaran. Metode ini sayangnya tidak dapat mendeteksi kecacatan visual pada permukaan dalam dan luar cincin bearing. Sebuah pengenalan pola diimplementasikan dalam paper ini untuk memecahkan masalah tersebut. Sebuah arsitektur jaringan saraf backpropagation digunakan untuk mengenali pola kecacatan visual. Arsitektur yang diusulkan ini terintegrasi dalam sebuah alir pengolahan citra digital. Tingkat pengenalan bearing yang baik adalah 92.93%, sedangkan untuk bantalan yang cacat adalah 75%. Angka ini menunjukkan integrasi jaringan syaraf tiruan dengan pengolahan citra digital dapat diterapkan untuk mendeteksi kecacatan visual pada bearing.

show abstract

A time delay neural network algorithm for estimating image-pattern shape and motion

Cited by 33 publications

References 23 publications

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

The Benefits of Dense Stereo for Pedestrian Detection

Untitled

Contact Info

Product

Resources

About