Tracking-by-Segmentation with Online Gradient Boosting Decision Tree

Son, Jae Sung; Jung, Ilchae; Park, Kayoung; Han, Bohyung

doi:10.1109/iccv.2015.350

Cited by 115 publications

(74 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because it does not require making assumptions on the data, it is extensively used in certain fields, such as in the optimization of recommendation systems [62,63], visual tracking algorithms [64], and traffic systems [65][66][67][68]. The attractiveness of GBRT comes from its ability to deal with the uneven distribution of data attributes, its lack of limitation for any hypothesis of input data, its better predictive capacity than a single decision tree, its power to deal with larger data size, and its transparency in terms of model development.…”

Section: Gradient Boosting Regression Treementioning

confidence: 99%

Estimating Surface Downward Shortwave Radiation over China Based on the Gradient Boosting Decision Tree Method

et al. 2018

View full text Add to dashboard Cite

Downward shortwave radiation (DSR) is an essential parameter in the terrestrial radiation budget and a necessary input for models of land-surface processes. Although several radiation products using satellite observations have been released, coarse spatial resolution and low accuracy limited their application. It is important to develop robust and accurate retrieval methods with higher spatial resolution. Machine learning methods may be powerful candidates for estimating the DSR from remotely sensed data because of their ability to perform adaptive, nonlinear data fitting. In this study, the gradient boosting regression tree (GBRT) was employed to retrieve DSR measurements with the ground observation data in China collected from the China Meteorological Administration (CMA) Meteorological Information Center and the satellite observations from the Advanced Very High Resolution Radiometer (AVHRR) at a spatial resolution of 5 km. The validation results of the DSR estimates based on the GBRT method in China at a daily time scale for clear sky conditions show an R 2 value of 0.82 and a root mean square error (RMSE) value of 27.71 W·m −2 (38.38%). These values are 0.64 and 42.97 W·m −2 (34.57%), respectively, for cloudy sky conditions. The monthly DSR estimates were also evaluated using ground measurements. The monthly DSR estimates have an overall R 2 value of 0.92 and an RMSE of 15.40 W·m −2 (12.93%). Comparison of the DSR estimates with the reanalyzed and retrieved DSR measurements from satellite observations showed that the estimated DSR is reasonably accurate but has a higher spatial resolution. Moreover, the proposed GBRT method has good scalability and is easy to apply to other parameter inversion problems by changing the parameters and training data.

show abstract

Section: Gradient Boosting Regression Treementioning

confidence: 99%

Estimating Surface Downward Shortwave Radiation over China Based on the Gradient Boosting Decision Tree Method

et al. 2018

View full text Add to dashboard Cite

show abstract

“…1) Non-rigid Object Tracking: The most recent dataset of non-rigid object tracking [14] includes 11 challenging image sequences with pixel-wise annotations in each frame. The [14]) and seven state-of-the-art bounding box-based trackers (Struck [70], SCM [71], MEEM [72], MUSTer [73], DSMT [28], FCNT [36], and HCFT [37]). These trackers based on hard-crafted [70]- [73] or deep [28], [36], [37] features achieve top performance on recent largescale tracking benchmarks [66].…”

Section: Experimental Results On Visual Trackingmentioning

confidence: 99%

“…Qualitative evaluation of the proposed tracker and other segmentation-based methods. From left to right: (a) input frames; (b) HT [13]; (c) SPT [11]; (d) PT [10]; (e) OGBDT [14]; (f) our method; and (g) ground truth. performance in dealing with traditional saliency detection and visual tracking problems.…”

Section: Discussionmentioning

confidence: 99%

“…Other datasets like JuddDB [52] and SED [67] are frequently used to evaluate saliency detection methods; however, we focus on the single salient object detection in this paper, which is considerably related to the online tracking problem. The recent tracking dataset presented in [14] is adopted to highlight the tracking process of non-rigid and articulated objects for the tracking task. In addition, the popular OTB-50 [66] dataset is used to evaluate the generalization ability of our tracker.…”

Section: A Datasetsmentioning

confidence: 99%

See 1 more Smart Citation

Non-rigid object tracking via deep multi-scale spatial-temporal discriminative saliency maps

Zhang

Liu

Wang

et al. 2020

Pattern Recognition

View full text Add to dashboard Cite

In this paper we propose an effective non-rigid object tracking method based on spatial-temporal consistent saliency detection. In contrast to most existing trackers that use a bounding box to specify the tracked target, the proposed method can extract the accurate regions of the target as tracking output, which achieves better description of the non-rigid objects while reduces background pollution to the target model. Furthermore, our model has several unique features. First, a tailored deep fully convolutional neural network (TFCN) is developed to model the local saliency prior for a given image region, which not only provides the pixel-wise outputs but also integrates the semantic information. Second, a multi-scale multi-region mechanism is proposed to generate local region saliency maps that effectively consider visual perceptions with different spatial layouts and scale variations. Subsequently, these saliency maps are fused via a weighted entropy method, resulting in a final discriminative saliency map. Finally, we present a non-rigid object tracking algorithm based on the proposed saliency detection method by utilizing a spatial-temporal consistent saliency map (STCSM) model to conduct target-background classification and using a simple fine-tuning scheme for online updating. Numerous experimental results demonstrate that the proposed algorithm achieves competitive performance in comparison with state-ofthe-art methods for both saliency detection and visual tracking, especially outperforming other related trackers on the non-rigid object tracking dataset.

show abstract

“…Gradient boosting is a highly matured algorithm to update the boosting classifiers online and frequently enhances the exactness of the segmentation [5]. In order to manage the labeling uncertainty for semi-supervised learning and it explains the online gradient boosting technique which depends on the various instance learning.…”

Section: Introductionmentioning

confidence: 99%

Improved Bat Algorithm using Super-Region Volume Segmentation for Medical Images

Nigam¹,

Ramanathaiah²,

B.³

2019

IJITEE

View full text Add to dashboard Cite

 Abstract: Medical imaging has been the highly active research area from the past few dec-ades. With the intention of consolidating the automatic approaches of power of the segmentation with the experienced and capable of annotating biological sam-ples manually, provides a new board which is termed as Super-Region Volume Segmentation (SuRVoS). SuRVoS purifies the estimation by utilizing Markov Random Field (MRF) formulation, which in turn considers neighboring labels in order to estimate precisely consistent. Nevertheless, in the MRF formulation model, various separations play a vital part in order to raise the exactness of out-put of the segmentation. Hence, identifying that partition number is an important task and also it is a difficult problem. In proposed work, the edge weight is seg-regated into two disjoints sets, which means cut and off into all feasible disjoint segments. The proposed energy function is reduced by utilizing Enhanced Bat (EB) algorithm over the MRF and the standard deviation calculation among two edges. In our work, the improved SuRVoS gives good segmentation output by generating good objective function values in the form of best supervoxels. So, it raises the accuracy of image segmentation greater instead of the earlier methods. The performance metrics were computed namely accuracy, precision, recall, Peak Signal to Noise Ratio (PSNR), Mean Square Error (MSE) and correlation values by utilizing the proposed method. The experimental output gives the greater seg-mentation accuracy which is accomplished by the proposed method.

show abstract

Tracking-by-Segmentation with Online Gradient Boosting Decision Tree

Cited by 115 publications

References 31 publications

Estimating Surface Downward Shortwave Radiation over China Based on the Gradient Boosting Decision Tree Method

Estimating Surface Downward Shortwave Radiation over China Based on the Gradient Boosting Decision Tree Method

Non-rigid object tracking via deep multi-scale spatial-temporal discriminative saliency maps

Improved Bat Algorithm using Super-Region Volume Segmentation for Medical Images

Contact Info

Product

Resources

About