Implicit hierarchical boosting for multi-view object detection

Perrotton, Xavier; Sturzel, Marc; Roux, Michel

doi:10.1109/cvpr.2010.5540115

Cited by 19 publications

(26 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some authors [10] [2] modified the loss function to use AdaBoost for regression. Perrotton et al [11] use Gentle Adaboost with a different families of descriptors and soft cascade structure [12]. They also modify the way of building the weak classifiers in order to build multi-view object detector.…”

Section: A State-of-the-artmentioning

confidence: 99%

Improving cascade of classifiers by sliding window alignment in between

Zimmermann

Hurych

Svoboda

2011

The 5th International Conference on Automation, Robotics and Applications

View full text Add to dashboard Cite

Abstract-We improve an object detector based on cascade of classifiers by a local alignment of the sliding window. The detector needs to operate on a relatively sparse grid in order to achieve a real time performance on high-resolution images. The proposed local alignment in the middle of the cascade improves its recognition performance whilst retaining the necessary speed. We show that the moment of the alignment matters and discuss the performance in terms of false negatives and false positives. The proposed method is tested on a car detection problem.

show abstract

Section: A State-of-the-artmentioning

confidence: 99%

Improving cascade of classifiers by sliding window alignment in between

Zimmermann

Hurych

Svoboda

2011

The 5th International Conference on Automation, Robotics and Applications

View full text Add to dashboard Cite

show abstract

“…When the training of current stage finished, we update the weight of samples according to the best performance over sub-categories corresponding to original target class: New weight of samples is used by implicit partition for next round. The decreasing structure has been observed as the best suited model in [6]. It ensures a faster fitting at the beginning of training meanwhile a better generalization capacity in following stages.…”

Section: Learning Hierarchical Structurementioning

confidence: 94%

“…We then deploy the JointBoost with sub-categories label set instead of original label set. Since every single partition is imperfect, we also employ the hierarchical structure as suggested in [6]. It has been proved that the hierarchical structure ensures an evolving weak partition according to sample weights.…”

Section: B Multiclass Implicit Partitionmentioning

confidence: 99%

“…We extend the implicit partition method in [6] to multiclass frameworks. The main idea of implicit partition is to loose the cost function for every target class from:…”

Section: B Multiclass Implicit Partitionmentioning

confidence: 99%

“…Second, a linear combination of such weak classifiers introduces an increasing obscurity with more intra-class variation. To solve these problems, [6] utilizes multi-layer perceptrons (MLP) as weak classifiers for the implicit partition of target samples. The weak classifiers on different partitions are then combined as a hierarchical cascade.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Implicit JointBoost for multiclass object detection under high intra-class variation

Zhang

Cheng

et al. 2012

2012 IEEE International Conference on Information Science and Technology

View full text Add to dashboard Cite

In this paper, we address the problem of using boosting to detect different classes of objects with significant intra-class variation. Current approaches generally require prior knowledge such as semantic information to obtain an explicit partition over positive samples. This can require a lot of manually labeled samples and can limit the performance of classifier due to subjective error. We present a novel JointBoost based learning method which can learn the variant appearances of targets without any prior assumption. With an implicit partition, the specific features for fractional samples can contribute to classification as well as generic features for the whole training set. By encouraging an intra-class and inter-class feature sharing between implicit sub-categories, our data-driven learning approach avoids a local optimum in the candidate weak classifier space. Experimental results on two popular tasks demonstrate the considerable improvements brought by the new approach. We hope that implicit JointBoost will extend the application of traditional boosting methods.

show abstract

Image Mosaic Technology Based on Harris Corner Feature

Liu

Wang

2021

Proceedings of the 11th International Conference on Computer Engineering and Networks

View full text Add to dashboard Cite

Implicit hierarchical boosting for multi-view object detection

Cited by 19 publications

References 20 publications

Improving cascade of classifiers by sliding window alignment in between

Improving cascade of classifiers by sliding window alignment in between

Implicit JointBoost for multiclass object detection under high intra-class variation

Image Mosaic Technology Based on Harris Corner Feature

Contact Info

Product

Resources

About