X-ray imaging system

A system comprising a processor with associated memory configured to perform operations including:

• obtaining an input image for processing based on a source X-ray image of a sample acquired using an X-ray detector;
• determining a model of the sample based on the input image by applying a model generation process to the input image, the model generation process comprising:
• obtaining a classification of one or more pixels of the image as corresponding to a given material composition, the classification based on comparison of intensity value(s) of the pixels to a threshold,
• determining an estimated material thickness value for each of the one or more classified pixels based on data defining expected thickness values for pixels having the given material composition,
• fitting a thickness function to the image based on estimated thickness values for the classified pixels, the thickness function indicating an estimated material thickness for each pixel of the image,
• estimating material composition values for each pixel based on the estimated material thickness values specified by the thickness function by determining for each pixel a set of non-implausible material composition values based on the material thickness value for the pixel and deriving the material composition value for the pixel from the set of non-implausible material values by averaging the set of non-implausible material values, and
• determining a model comprising for each pixel a material composition value and a thickness value based on the estimated material composition values and the estimated material thickness values;
• evaluating the model by a process comprising:
• computing, based on the model, simulated X-ray image data;
• evaluating the simulated image data against the input image to determine whether a convergence criterion is fulfilled;
• and
• generating an updated model of the sample if the convergence criterion is not fulfilled by computing a scatter-corrected version of the input image using the simulated image data by removing a simulated scatter image from the input image, and applying the model generation process to the scatter-corrected input image;
• the operations further including:
• repeating the model evaluating process based on one or more successive updated models until the convergence criterion is fulfilled in a final iteration;
• and
• performing scatter correction on the source X-ray image using simulated image data computed during the final iteration.

A non-transitory computer-readable medium comprising software code which when executed is arranged to perform a method for scatter correction of an X-ray image, comprising:

• obtaining an input image for processing based on a source X-ray image of a sample acquired using an X-ray detector;
• determining a model of the sample based on the input image by applying a model generation process to the input image, the model generation process comprising:
• obtaining a classification of one or more pixels of the image as corresponding to a given material composition, the classification based on comparison of intensity value(s) of the pixels to a threshold,
• determining an estimated material thickness value for each of the one or more classified pixels based on data defining expected thickness values for pixels having the given material composition,
• fitting a thickness function to the image based on estimated thickness values for the classified pixels, the thickness function indicating an estimated material thickness for each pixel of the image,
• estimating material composition values for each pixel based on the estimated material thickness values specified by the thickness function by determining for each pixel a set of non-implausible material composition values based on the material thickness value for the pixel and deriving the material composition value for the pixel from the set of non-implausible material values by averaging the set of non-implausible material values, and
• determining a model comprising for each pixel a material composition value and a thickness value based on the estimated material composition values and the estimated material thickness values;
• evaluating the model by a process comprising:
• computing, based on the model, simulated X-ray image data;
• evaluating the simulated image data against the input image to determine whether a convergence criterion is fulfilled;
• and
• generating an updated model of the sample if the convergence criterion is not fulfilled by computing a scatter-corrected version of the input image using the simulated image data by removing a simulated scatter image from the input image, and applying the model generation process to the scatter-corrected input image;
• the method comprising repeating the model evaluating process based on one or more successive updated models until the convergence criterion is fulfilled in a final iteration;
• and
• performing scatter correction on the source X-ray image using simulated image data computed during the final iteration.