Triplet Loss in Siamese Network for Object Tracking

Dong, Xingping; Shen, Jianbing

doi:10.1007/978-3-030-01261-8_28

Cited by 503 publications

(235 citation statements)

References 33 publications

Supporting

Mentioning

232

Contrasting

Unclassified

Order By: Relevance

“…Recently, trackers based on Siamese networks [2,7,12,13,20,34,40] have drawn great attention due to their high speed and accuracy. However, the backbone network utilized in these trackers is still the classical AlexNet [18], * corresponding author Here, width refers to the number of branches in a module.…”

Section: Introductionmentioning

confidence: 99%

Deeper and Wider Siamese Networks for Real-Time Visual Tracking

Zhang

Peng

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

918

491

View full text Add to dashboard Cite

Siamese networks have drawn great attention in visual tracking because of their balanced accuracy and speed. However, the backbone networks used in Siamese trackers are relatively shallow, such as AlexNet [18], which does not fully take advantage of the capability of modern deep neural networks. In this paper, we investigate how to leverage deeper and wider convolutional neural networks to enhance tracking robustness and accuracy. We observe that direct replacement of backbones with existing powerful architectures, such as ResNet [14] and Inception [33], does not bring improvements. The main reasons are that 1) large increases in the receptive field of neurons lead to reduced feature discriminability and localization precision; and 2) the network padding for convolutions induces a positional bias in learning. To address these issues, we propose new residual modules to eliminate the negative impact of padding, and further design new architectures using these modules with controlled receptive field size and network stride. The designed architectures are lightweight and guarantee real-time tracking speed when applied to SiamFC [2] and SiamRPN [20]. Experiments show that solely due to the proposed network architectures, our SiamFC+ and SiamRPN+ obtain up to 9.8%/5.7% (AUC), 23.3%/8.8% (EAO) and 24.4%/25.0% (EAO) relative improvements over the original versions [2, 20] on the OTB-15, VOT-16 and VOT-17 datasets, respectively.

show abstract

Section: Introductionmentioning

confidence: 99%

Deeper and Wider Siamese Networks for Real-Time Visual Tracking

Zhang

Peng

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

918

491

View full text Add to dashboard Cite

show abstract

“…They proposed a fully-convolutional Siamese network and achieved state-1 https://github.com/SilvioGiancola/ShapeCompletion3DTracking of-the-art performance on the VOT benchmark [28]. Recent Siamese trackers estimate boundary flows [31], use contextual structure [20], attention [57], distraction [65], semantic information [63], triplet losses [14] and region proposal networks [32] to improve tracking performance. To the best of our knowledge, our work is the first 3D adaptation of Siamese networks for 3D point cloud tracking.…”

Section: Related Workmentioning

confidence: 99%

Leveraging Shape Completion for 3D Siamese Tracking

Giancola

Zarzar

Ghanem

2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

154

207

View full text Add to dashboard Cite

Point clouds are challenging to process due to their sparsity, therefore autonomous vehicles rely more on appearance attributes than pure geometric features. However, 3D LIDAR perception can provide crucial information for urban navigation in challenging light or weather conditions. In this paper, we investigate the versatility of Shape Completion for 3D Object Tracking in LIDAR point clouds. We design a Siamese tracker that encodes model and candidate shapes into a compact latent representation. We regularize the encoding by enforcing the latent representation to decode into an object model shape. We observe that 3D object tracking and 3D shape completion complement each other. Learning a more meaningful latent representation shows better discriminatory capabilities, leading to improved tracking performance. We test our method on the KITTI Tracking set using car 3D bounding boxes. Our model reaches a 76.94% Success rate and 81.38% Precision for 3D Object Tracking, with the shape completion regularization leading to an improvement of 3% in both metrics.

show abstract

“…For comprehensive evaluation and analysis on the effectiveness of our approach, we compare SATIN with five state-of-the-art trackers, including SiamRPN [25], CREST [61], SiamTri [62], ACFN [60], and SiamFC [21] in different attributes on the OTB-2015 benchmark. Each sequence in the OTB-2015 benchmark dataset is categorized with 11 attributes.…”

Section: Success Performance On Otb-2015 With 11 Attributes (100)mentioning

confidence: 99%

Siamese attentional keypoint network for high performance visual tracking

Gao

Yuan

Wang

et al. 2020

Knowledge-Based Systems

View full text Add to dashboard Cite

Visual tracking is one of the most fundamental topics in computer vision. Numerous tracking approaches based on discriminative correlation filters or Siamese convolutional networks have attained remarkable performance over the past decade. However, it is still commonly recognized as an open research problem to develop robust and effective trackers which can achieve satisfying performance with high computational and memory storage efficiency in real-world scenarios. In this paper, we investigate the impacts of three main aspects of visual tracking, i.e., the backbone network, the attentional mechanism, and the detection component, and propose a Siamese Attentional Keypoint Network, dubbed SATIN, for efficient tracking and accurate localization. Firstly, a new Siamese lightweight hourglass network is specially designed for visual tracking. It takes advantage of the benefits of the repeated bottom-up and top-down inference to capture more global and local contextual information at multiple scales. Secondly, a novel cross-attentional module is utilized to leverage both channel-wise and spatial intermediate attentional information, which can enhance both discriminative and localization capabilities of feature maps. Thirdly, a keypoints detection approach is invented to trace any target object by detecting the top-left corner point, the centroid point, and the bottom-right corner point of its bounding box. Therefore, our SATIN tracker not only has a strong capability to learn more effective object representations, but also is computational and memory storage efficiency, either during the training or testing stages. To the best of our knowledge, we are the first to propose this approach. Without bells and whistles, experimental results demonstrate that our approach achieves state-of-the-art performance on several recent benchmark datasets, at a speed far exceeding 27 frames per second. with real-time speed are widely pursued in numerous applications, including video surveillance, autonomous driving, pose estimation, and human-computer interfaces. Although considerable progress has been made by many tracking approaches in the last decade, visual tracking is still a challenging problem due to multiple negative scenarios such as occlusions, fast motions, scale variations, and background clutters [1,2,3,4].Most recent tracking approaches are developed based on two successful frameworks. The first one is the discriminative correlation filter (DCF). Thanks to the circular shifting of training samples and fast learning correlation filters in the Fourier frequency domain, DCF demonstrates superior computational efficiency and fairly good tracking accuracy and has attracted increasing attention since MOSSE [5]first exploited it. Later, in order to overcome different kinds of challenges and achieve competitive tracking performance, the advancement in DCF-based tracking is focused on the use of kernel functions [6], motion information [7,8], multi-dimensional features [9,10], multi-scale estimation [11,12], boundary effects allevi...

show abstract

Triplet Loss in Siamese Network for Object Tracking

Cited by 503 publications

References 33 publications

Deeper and Wider Siamese Networks for Real-Time Visual Tracking

Deeper and Wider Siamese Networks for Real-Time Visual Tracking

Leveraging Shape Completion for 3D Siamese Tracking

Siamese attentional keypoint network for high performance visual tracking

Contact Info

Product

Resources

About