# Real-Time Correlation Tracking via Joint Model Compression and Transfer

arXiv: https://arxiv.org/abs/1907.09831 · HTML: https://mlsanity.com/paper/1907.09831/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Jul 2019.

## Built on by (most cited first, 8 of 8)

- [Event Stream-based Visual Object Tracking: A High-Resolution Benchmark Dataset and A Novel Baseline](https://mlsanity.com/paper/2309.14611/) (2023)
- [All-Day Object Tracking for Unmanned Aerial Vehicle](https://mlsanity.com/paper/2101.08446/) (2021)
- [Predictive Visual Tracking: A New Benchmark and Baseline Approach](https://mlsanity.com/paper/2103.04508/) (2021)
- [Towards Feature Distribution Alignment and Diversity Enhancement for Data-Free Quantization](https://mlsanity.com/paper/2205.00179/) (2022)
- [Event Stream-based Visual Object Tracking: HDETrack V2 and A High-Definition Benchmark](https://mlsanity.com/paper/2502.05574/) (2025)
- [MVTD: A Benchmark Dataset for Maritime Visual Object Tracking](https://mlsanity.com/paper/2506.02866/) (2025)
- [Altering Backward Pass Gradients improves Convergence](https://mlsanity.com/paper/2111.12495/) (2021)
- [Infrared UAV Target Tracking with Dynamic Feature Refinement and Global Contextual Attention Knowledge Distillation](https://mlsanity.com/paper/2512.04581/) (2025)

## Abstract (arXiv)

Correlation filters (CF) have received considerable attention in visual tracking because of their computational efficiency. Leveraging deep features via off-the-shelf CNN models (e.g., VGG), CF trackers achieve state-of-the-art performance while consuming a large number of computing resources. This limits deep CF trackers to be deployed to many mobile platforms on which only a single-core CPU is available. In this paper, we propose to jointly compress and transfer off-the-shelf CNN models within a knowledge distillation framework. We formulate a CNN model pretrained from the image classification task as a teacher network, and distill this teacher network into a lightweight student network as the feature extractor to speed up CF trackers. In the distillation process, we propose a fidelity loss to enable the student network to maintain the representation capability of the teacher network. Meanwhile, we design a tracking loss to adapt the objective of the student network from object recognition to visual tracking. The distillation process is performed offline on multiple layers and adaptively updates the student network using a background-aware online learning scheme. Extensive experiments on five challenging datasets demonstrate that the lightweight student network accelerates the speed of state-of-the-art deep CF trackers to real-time on a single-core CPU while maintaining almost the same tracking accuracy.