This paper presents a novel discriminative scale space tracking approach for visual object tracking. The method improves upon existing state-of-the-art tracking methods by learning separate discriminative correlation filters for translation and scale estimation. The scale filter is learned online using target appearance samples at different scales, allowing for accurate and robust scale estimation. The method is computationally efficient, operating at a 50% higher frame rate compared to standard exhaustive scale search methods. Extensive experiments on the OTB and VOT2014 datasets show that the proposed approach achieves a 2.5% improvement in average overlap precision on OTB and outperforms 19 state-of-the-art trackers on OTB and 37 on VOT2014. The method is also shown to be robust to scale variations and provides significant improvements in tracking performance. The approach is based on learning multi-channel discriminative correlation filters, which are used to estimate both translation and scale. The method is further enhanced by reducing computational cost through strategies such as sub-grid interpolation and dimensionality reduction. The resulting fast discriminative scale space tracker (fDSST) achieves significantly improved tracking performance while operating at twice the speed of the standard DSST method. The experiments show that the proposed approach significantly improves the performance of standard translation trackers, demonstrating the importance of accurate scale estimation for robust tracking. The method is also shown to be effective in real-time applications, with the fDSST achieving a 7.0% and 4.4% improvement in mean OP and DP respectively on the OTB dataset. The results demonstrate that the proposed approach is both accurate and efficient, making it suitable for a wide range of real-time tracking applications.This paper presents a novel discriminative scale space tracking approach for visual object tracking. The method improves upon existing state-of-the-art tracking methods by learning separate discriminative correlation filters for translation and scale estimation. The scale filter is learned online using target appearance samples at different scales, allowing for accurate and robust scale estimation. The method is computationally efficient, operating at a 50% higher frame rate compared to standard exhaustive scale search methods. Extensive experiments on the OTB and VOT2014 datasets show that the proposed approach achieves a 2.5% improvement in average overlap precision on OTB and outperforms 19 state-of-the-art trackers on OTB and 37 on VOT2014. The method is also shown to be robust to scale variations and provides significant improvements in tracking performance. The approach is based on learning multi-channel discriminative correlation filters, which are used to estimate both translation and scale. The method is further enhanced by reducing computational cost through strategies such as sub-grid interpolation and dimensionality reduction. The resulting fast discriminative scale space tracker (fDSST) achieves significantly improved tracking performance while operating at twice the speed of the standard DSST method. The experiments show that the proposed approach significantly improves the performance of standard translation trackers, demonstrating the importance of accurate scale estimation for robust tracking. The method is also shown to be effective in real-time applications, with the fDSST achieving a 7.0% and 4.4% improvement in mean OP and DP respectively on the OTB dataset. The results demonstrate that the proposed approach is both accurate and efficient, making it suitable for a wide range of real-time tracking applications.