Transformers in Single Object Tracking: An Experimental Survey

Thangavel, Janani; Thanikasalam, Kokul; Ramanan, Amirthalingam; Fernando, Subha

doi:10.48550/arxiv.2302.11867

Cited by 4 publications

(6 citation statements)

References 93 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the large number of matrix multiplication computations required by the attention mechanism used dramatically slows down the inference while improving the performance; as in, for example, the recent SeqTrack [14] and MixFormer [15], which, although they achieved 72.5% and 70.0% success rates with the LaSOT dataset, respectively, utilized 535.85 G and 113.02 G model flops and 308.98 M and 195.40 M parameters for the best results. At the same time, their speeds were only 5.81 and 8.02 FPS [16].…”

Section: Introductionmentioning

confidence: 92%

See 1 more Smart Citation

Scaling-Invariant Max-Filtering Enhancement Transformers for Efficient Visual Tracking

Chen,

Xiong,

Meng

et al. 2023

Preprint

View full text Add to dashboard Cite

Real-time tracking is one of the most challenging problems in computer vision. Most Transformer-based trackers usually require expensive computational and storage power, which leads to the inability of these robust trackers to achieve real-time performance in resource-constrained devices. In this work, we propose a lightweight tracker, AnteaTrack. To localize the target more precisely, we develop a scaling-invariant max-filtering operator employing a sliding window combined with local max-pooling, which filters out the suspected target from the feature and performs an augmented representation while suppressing the background. For a more tight bounding box, we employ Pixel-Shuffle to increase the resolution of the feature map and get a more fine-grained representation of the target. In addition, AnteaTrack can run in real-time at 47 frames per second(FPS) on the CPU. We tested AnteaTrack on 5 datasets, and a large number of experiments have shown that AnteaTrack provides the most efficient solution compared to the same type of CPU real-time trackers. The code will be available at https://github.com/cnchange/AnteaTrack.

show abstract

Section: Introductionmentioning

confidence: 92%

“…Recently, Yan [33] introduced Transformer into target tracking and achieved good performance. Transformer-based tracking can be divided into two categories [16]. One is CNN Transformer-based trackers [33][34][35][36].…”

Section: Transformer-based Trackermentioning

confidence: 99%

Scaling-Invariant Max-Filtering Enhancement Transformers for Efficient Visual Tracking

Chen,

Xiong,

Meng

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…And fully-transformer based trackers like SwinTrack [27], MixFormer [19], Pro-ContEXT [28] and VideoTrack [29]. The state-ofthe-art tracker tends to utilize a fully-transformer architecture as it achieves the best performance [30], while CNN transformer architecture offers a balanced performance between speed and accuracy [30].…”

Section: Transformer-based Trackersmentioning

confidence: 99%

Enhancing Single Object Tracking with Temporal Information

Abboud,

Jafar,

Hamdoun

2024

Preprint

View full text Add to dashboard Cite

This study introduces an enhancement approach for single object tracking (SOT) algorithms, aimingto use temporal information with minimize architectural changes while preserving the original modeldimensions with minimal impact on the speed. Most state-of-the-art approaches prioritize effectivefeature extraction from the object or the search window images, while temporal information is oftenneglected. Additionally, many deep models used in SOTs are time-consuming compared to traditionaltrackers. This study focuses on SwinTrack V2 as our case study due to its absence of temporalinformation utilization and its excellent performance on GOT-10k dataset. The method suggest addinga third input to the model, which is an updated object template. The initial template features arefused with the updated template using an attention block, which effectively preserves the originalmodel dimension. Our additional block improves performance when testing on Got-10k dataset by 0.9% for average overlap AO and by 2.1% for success rate SR0.5 compared to the original tracker,with only a slight impact on speed.

show abstract

“…According to the type of the framework, the SOT algorithms using Transformer can be classified as CNN-Transformer based trackers and fully-Transformer based trackers (Thangavel et al 2023).…”

Section: Transformer-based Trackersmentioning

confidence: 99%

Sequential Fusion Based Multi-Granularity Consistency for Space-Time Transformer Tracking

Hu,

Yang,

Huang

et al. 2024

AAAI

View full text Add to dashboard Cite

Regarded as a template-matching task for a long time, visual object tracking has witnessed significant progress in space-wise exploration. However, since tracking is performed on videos with substantial time-wise information, it is important to simultaneously mine the temporal contexts which have not yet been deeply explored. Previous supervised works mostly consider template reform as the breakthrough point, but they are often limited by additional computational burdens or the quality of chosen templates. To address this issue, we propose a Space-Time Consistent Transformer Tracker (STCFormer), which uses a sequential fusion framework with multi-granularity consistency constraints to learn spatiotemporal context information. We design a sequential fusion framework that recombines template and search images based on tracking results from chronological frames, fusing updated tracking states in training. To further overcome the over-reliance on the fixed template without increasing computational complexity, we design three space-time consistent constraints: Label Consistency Loss (LCL) for label-level consistency, Attention Consistency Loss (ACL) for patch-level ROI consistency, and Semantic Consistency Loss (SCL) for feature-level semantic consistency. Specifically, in ACL and SCL, the label information is used to constrain the attention and feature consistency of the target and the background, respectively, to avoid mutual interference. Extensive experiments have shown that our STCFormer outperforms many of the best-performing trackers on several popular benchmarks.

show abstract

Transformers in Single Object Tracking: An Experimental Survey

Cited by 4 publications

References 93 publications

Scaling-Invariant Max-Filtering Enhancement Transformers for Efficient Visual Tracking

Scaling-Invariant Max-Filtering Enhancement Transformers for Efficient Visual Tracking

Enhancing Single Object Tracking with Temporal Information

Sequential Fusion Based Multi-Granularity Consistency for Space-Time Transformer Tracking

Contact Info

Product

Resources

About