Online Object Tracking: A Benchmark

By Yi Wu, Jongwoo Lim, Ming-Hsuan Yang. 2013

3,975 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Simitate: A Hybrid Imitation Learning Benchmark (2019, neutral): “many research topics like autonomous driving [14], RGB-D SLAM systems [15], object tracking [13], [37], [38].”
Learning Multi-domain Convolutional Neural Networks for Visual Tracking (2015, background): “(b) OTB100 result Figure 3: Precision and success plots on OTB50 [41] and OTB100 [40].”
Fully-Convolutional Siamese Networks for Object Tracking (2016, methodology): “Success plots for OPE (one pass evaluation), TRE (temporal robustness evaluation) and SRE (spatial robustness evaluation) of the OTB-13 [11] benchmark.”
Unsupervised Object Discovery and Tracking in Video Collections (2015, background): “Conventional object tracking methods [41] usually require annotations for at least one frame [13, 15, 40], or object detectors trained for target classes in a supervised manner [1, 2, 26].”
Transferring Rich Feature Hierarchies for Robust Visual Tracking (2015, methodology): “Recent work [33] introduced a uniﬁed tracking benchmark to reduce the bias.”
Multiple object tracking: A literature review (2014, background): “• The third set [29], [30] introduces and discusses benchmarks on general visual tracking [29] and on specific multiple object tracking [30].”
Deep Learning for Visual Tracking: A Comprehensive Survey (2019, methodology): “The OTB and LaSOT toolkits evaluate the visual tracking methods in terms of the well-known precision and success plots and then rank the methods based on AUC score [185], [186].”
Accurate Scale Estimation for Robust Visual Tracking (2014, methodology): “In addition, the results are presented using precision and success plots [18].”
Struck: Structured Output Tracking with Kernels (2016, neutral): “[21] benchmark and compare them to state-of-the-art trackers, and in Section 7 we conclude.”
Siam R-CNN: Visual Tracking by Re-Detection (2019, positive): “We evaluate on OTB2015 [70] (100 videos, 590 frames average length), calculating the success and precision over varying overlap thresholds.”
The Visual Object Tracking VOT2014 Challenge Results (2014, positive): “OTB [62] evaluates trackers using two measures: precision score and success score.”
Convolutional Features for Correlation Filter Based Visual Tracking (2015, methodology): “formance on benchmark tracking datasets [24, 39].”
Hedged Deep Tracking (2016, methodology): “To evaluate the effectiveness of the proposed adaptive Hedge method, we compare the HDT with its constituent CNN trackers denoted by VGG-10, VGG-11, VGG-12, VGG-14, VGG-15, and VGG-16, as well as the hedged CNN tracker using the standard parameter-free Hedge [5], denoted by HDT-SH, on the benchmark ”
Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model (2017, methodology): “Benchmark for video tasks There exist benchmarks for several video tasks like online object tracking [31] and video object segmentation [23].”
Learning Background-Aware Correlation Filters for Visual Tracking (2017, methodology): “More particularly, BACF achieved the best AUC (67.78) on OTB50 followed by LCT (62.48) and SRDCF (62.35).”
Template Matching with Deformable Diversity Similarity (2016, methodology): “Our experimental setup follows that of [18] that created a benchmark by sampling frames from video sequences annotated with bounding-boxes for object tracking [28].”
ECO: Efficient Convolution Operators for Tracking (2016, methodology): “The trackers are evaluated using success plot [36], calculated as percentage of frames with an intersection-over-union (IOU) overlap exceeding a threshold.”
In defense of color-based model-free tracking (2015, methodology): “We compare our approach to recent state-of-the-art algorithms including the winner of the VOT14 challenge, DSST [13], and two of the top-performing trackers of the online tracking benchmark [44], namely Struck [21] and CSK [22].”
Understanding and Diagnosing Visual Tracking Systems (2015, methodology): “A milestone is the recent contribution made by a benchmark [39] which consists of 50 videos with full annotations.”