Worked on enhancing the DirectShow VideoCapture backend in the espressif/opencv repository, focusing on reducing camera startup latency and improving overall performance. Developed a new parameterized initialization approach by introducing the VideoCaptureParameters constructor, allowing pre-setting of device FourCC, width, and height during initialization. This change enabled faster camera open times and a more responsive video capture pipeline. The work was implemented in C++ with an emphasis on performance optimization and video capture reliability. No bugs were fixed during this period, but the feature addressed a specific bottleneck in camera initialization, demonstrating depth in targeted performance engineering within the codebase.