Computer-implemented method for determining the scatter radiation contribution to an evaluation image, evaluation device, computer program and electronically readable data carrier
DE102025105314B3Undetermined Publication Date: 2026-07-09SIEMENS HEALTHINEERS AG
Patent Information
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- SIEMENS HEALTHINEERS AG
- Filing Date
- 2025-02-13
- Publication Date
- 2026-07-09
Smart Images

Figure 00000000_0000_ABST
Abstract
The invention relates to a computer-implemented method for determining the scatter radiation contribution to an evaluation image (47) acquired with an X-ray device (57), comprising the following steps: - providing at least one encoder (35) of a trained detection function (34), which includes the encoder (35) that provides a feature vector (38) in a latent space for an X-ray image provided as input data, and a decoder (36) that determines an output image (39) describing the scatter radiation contribution in the X-ray image from the feature vector (38); - providing an evaluation database (33) in which reference data sets (45) are stored, which for each training data set with which the detection function (34) was trained each comprise at least: - the training scatter radiation image (40) of the training data set indicating the basic truth, and - the feature vector (38) for the training data set.- Applying the encoder (35) to the evaluation image (47) to determine a feature vector (38) of the evaluation image (47), - Comparing the feature vector (38) of the evaluation image (47) with the feature vectors (38) of the reference data sets (45) to determine at least one reference data set (45) with the most similar feature vector (38) from the evaluation database (44), and - Determining a scatter radiation image (46) for the evaluation image (47) from the training scatter radiation images (40) of the at least one determined reference data set (45).
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
DE102019215242A1