This paper presents a novel approach for robust object tracking using Multiple Instance Learning (MIL) in an online manner. The goal is to track an object in a video given its initial location without prior knowledge of its appearance. Traditional tracking methods often struggle with appearance changes due to occlusions, lighting variations, or deformation. The proposed method uses MIL to train a discriminative classifier that can handle ambiguities in the training data, leading to more robust tracking with fewer parameter adjustments. The key contribution is the development of an online MIL algorithm for object tracking, which outperforms existing methods in terms of robustness and stability. The algorithm uses a discriminative appearance model that is updated online based on positive and negative examples extracted from the current frame. The model is trained using a boosting framework, and the algorithm is designed to handle the challenges of tracking objects with varying scales and poses. The paper presents extensive experimental results on challenging video sequences, demonstrating that the proposed method achieves superior performance compared to other state-of-the-art tracking algorithms. The results show that the MIL-based tracker is more robust to partial occlusions and various appearance changes. The method is also extended to track both location and scale of objects, showing improved performance in scenarios with significant appearance changes. The paper also discusses the challenges of adaptive appearance models and highlights the advantages of using MIL in tracking scenarios where the object's appearance changes significantly. The proposed method is evaluated using both qualitative and quantitative metrics, including center location error and precision plots. The results show that the MIL-based tracker outperforms other methods in terms of accuracy and robustness. The paper concludes that using MIL for object tracking leads to more robust and stable tracking compared to traditional methods. The proposed algorithm is effective in handling challenging tracking scenarios and is suitable for real-time applications. The method is also applicable to other tracking problems, such as tracking multiple objects, contours, or deformable objects. The paper provides a comprehensive review of the state-of-the-art in adaptive appearance models and discusses the potential of online MIL in other computer vision tasks.This paper presents a novel approach for robust object tracking using Multiple Instance Learning (MIL) in an online manner. The goal is to track an object in a video given its initial location without prior knowledge of its appearance. Traditional tracking methods often struggle with appearance changes due to occlusions, lighting variations, or deformation. The proposed method uses MIL to train a discriminative classifier that can handle ambiguities in the training data, leading to more robust tracking with fewer parameter adjustments. The key contribution is the development of an online MIL algorithm for object tracking, which outperforms existing methods in terms of robustness and stability. The algorithm uses a discriminative appearance model that is updated online based on positive and negative examples extracted from the current frame. The model is trained using a boosting framework, and the algorithm is designed to handle the challenges of tracking objects with varying scales and poses. The paper presents extensive experimental results on challenging video sequences, demonstrating that the proposed method achieves superior performance compared to other state-of-the-art tracking algorithms. The results show that the MIL-based tracker is more robust to partial occlusions and various appearance changes. The method is also extended to track both location and scale of objects, showing improved performance in scenarios with significant appearance changes. The paper also discusses the challenges of adaptive appearance models and highlights the advantages of using MIL in tracking scenarios where the object's appearance changes significantly. The proposed method is evaluated using both qualitative and quantitative metrics, including center location error and precision plots. The results show that the MIL-based tracker outperforms other methods in terms of accuracy and robustness. The paper concludes that using MIL for object tracking leads to more robust and stable tracking compared to traditional methods. The proposed algorithm is effective in handling challenging tracking scenarios and is suitable for real-time applications. The method is also applicable to other tracking problems, such as tracking multiple objects, contours, or deformable objects. The paper provides a comprehensive review of the state-of-the-art in adaptive appearance models and discusses the potential of online MIL in other computer vision tasks.