Low-Dimensional Color Attribute Tracking with Scale-Adaptive Correlation Filters: Applications in Biotechnology Imaging and Diagnostics
DOI:
https://doi.org/10.5912/jcb1881Abstract
In biotechnology imaging and diagnostics, accurate and stable target tracking is critical for tissue analysis, cellular imaging, and diagnostic monitoring applications. However, challenges such as occlusion, non-rigid deformation, and illumination variations often cause tracking drift, reducing the reliability of traditional methods. To address these issues, this study proposes an improved target tracking algorithm that integrates color and texture features to enhance performance in complex environments. Building on the traditional kernel correlation filter approach, the algorithm extracts complementary gradient features from grayscale images and texture features from color images. The proposed method achieves more robust and precise target tracking by combining these features with a scale-adaptive strategy. The scale-adaptive mechanism dynamically adjusts to target size and appearance variations, ensuring continuity and accuracy in tracking over time. Experimental validation demonstrates the algorithm's strong robustness under multiple challenging conditions, including occlusion, deformation, and varying illumination. The results highlight its ability to track targets accurately and stably in real-time scenarios, making it highly applicable for biotechnology use cases. This approach provides a reliable tool for improving biotech research and healthcare imaging and diagnostics, enabling more precise analysis and monitoring in critical applications.