Worked on a performance-focused refactor of the Hough Transform Accumulator in the opencv/opencv repository, targeting faster Hough line detection without altering core algorithmic results. The approach introduced multi-threading using C++ and parallel programming techniques, specifically partitioning angle ranges to enable conflict-free writes and leveraging cv::parallel_for_ for efficient execution. Image processing expertise was applied to pre-scan edge points into contiguous memory, improving cache locality and reducing allocations. Extensive validation and benchmarking demonstrated up to 88% faster CPU performance in standard tests. The contribution included comprehensive documentation, tests, and build notes, aligning with OpenCV’s standards for production-ready adoption.