Developed a pixel sampling pipeline enhancement for the nerfstudio repository, focusing on efficient batch processing of variable-resolution image data. Refactored the sampling logic using Python to incorporate a defaultdict-based approach, which improved the robustness and organization of sampled pixel storage. This solution ensured accurate sampling and collation of both images and depth images across diverse input resolutions, directly addressing challenges in data integrity and scalability. Leveraging skills in computer vision and data processing, the work laid a foundation for more reliable and performant batch workflows, enabling nerfstudio to handle heterogeneous datasets with greater efficiency and maintainability in future development.