This paper introduces a pipeline for generating synthetic Inertial Measurement Unit data to address data scarcity in Human Activity Recognition (HAR). The ability to differentiate between activities benefits fall detection, fitness tracking, and health analysis. However, limited availability and high cost of labeled training data hinders robust model development. While current synthetic data approaches remain cumbersome and require extensive domain expertise, this automated pipeline makes data generation more accessible, enabling researchers to develop more powerful HAR systems. The proposed methodological framework combines video generation models with pose estimation techniques and physics-based kinematic transformations to produce synthetic sensor data that effectively augments real-world measurements. Through evaluation across established benchmark datasets, results demonstrate that models trained on a combination of real and synthetic data achieve improved performance compared to baseline approaches, with F1-score improvements of up to 1.7% points. The proposed pipeline offers several advantages, including cost-efficiency, accessibility on standard computational infrastructure, and privacy preservation capabilities that address key limitations of traditional data collection protocols. The consistent improvements observed across multiple experimental configurations demonstrate the potential of synthetic data generation for HAR applications, particularly for enhancing model robustness and improving recognition of rare or underrepresented activities.