Automatic Winston–Lutz Analysis in Python (Single EPID Image)

In this video, I demonstrate how to automatically analyze a single Winston–Lutz EPID image using Python. Starting from a DICOM image, we automatically: ✔ Read the DICOM image ✔ Detect the radiation field ✔ Locate the ball bearing (BB) using 2D Gaussian fitting ✔ Calculate the Winston–Lutz offset ✔ Report the result in IEC coordinates (dU and dV) This example uses a synthetic EPID image with a known ground truth, making it ideal for validating the image processing pipeline while avoiding the use of clinical data. Topics covered: • DICOM image processing • pydicom • OpenCV contour detection • Gaussian fitting for subpixel localization • Radiation field detection • Winston–Lutz QA • Medical Physics automation • Image processing in radiotherapy This is the first video in a Winston–Lutz analysis series. In upcoming videos, we'll extend this project to full 3D Winston–Lutz analysis by incorporating: • Gantry angle • Collimator angle • Couch angle • Multiple EPID images • Radiation isocenter reconstruction • Machine walkout characterization If you're interested in Medical Physics, Python programming, Radiotherapy QA, DICOM, or automation, consider subscribing for future tutorials. Thanks for watching! 📥 Get the complete Winston–Lutz QA Python Toolkit here: https://payhip.com/b/gyZ9X Includes the synthetic DICOM generator, automated WL image analyzer, and 3D isocenter reconstruction tools demonstrated in this video.