Person re-identification (ReID) is important in surveillance for identifying and tracking individuals, but challenges like pose variation, lighting changes, and occlusion persist. This study introduces a fusion-based framework combining deep features with traditional color-based features for a robust representation. Deep neural networks extract semantic features, while color-based features address appearance variations. By leveraging these complementary strengths, the method effectively manages both sequential and non-sequential image frames, offering versatility and improved accuracy across diverse surveillance scenarios. Extensive experiments on five datasets show that our framework surpasses state-of-the-art methods, highlighting fusion-based techniques’ potential to enhance ReID systems with improved accuracy, robustness, and reliability under diverse surveillance conditions.