초록 ▼

A system and method for detecting poor quality images in an optical tomography system includes an acquisition apparatus for acquiring a set of pseudo-projection images of an object having a center of mass, where each of the set of pseudo-projection images is acquired at a different angle of view. A

A system and method for detecting poor quality images in an optical tomography system includes an acquisition apparatus for acquiring a set of pseudo-projection images of an object having a center of mass, where each of the set of pseudo-projection images is acquired at a different angle of view. A reconstruction apparatus is coupled to receive the pseudo-projection images, for reconstruction of the pseudo-projection images into 3D reconstruction images. A quality apparatus is coupled to receive the 3D reconstruction images and operates to detect of selected features that characterize poor quality reconstructions.

대표청구항 ▼

1. A system for detecting poor quality images in an optical tomography system comprising: acquisition means for acquiring a set of pseudo-projection images of an object having a center of mass, where each of the set of pseudo-projection images is acquired at a different angle of view;reconstruction

1. A system for detecting poor quality images in an optical tomography system comprising: acquisition means for acquiring a set of pseudo-projection images of an object having a center of mass, where each of the set of pseudo-projection images is acquired at a different angle of view;reconstruction means, coupled to receive the pseudo-projection images, for reconstruction of the pseudo-projection images into 3D reconstruction images; andquality means for classification of the 3D reconstruction images using selected features that characterize poor quality reconstructions wherein the set of pseudo-projection images present a center of mass trend for the object and the selected features are calculated from a comparison of the center of mass trend with a cosine function fitting the trend in the center of mass. 2. The system of claim 1 wherein the selected features describe streaking in the 3D reconstruction images. 3. The system of claim 1 wherein the selected features include difference image features that are formed by creating a difference image between a fixed focus image and a reconstruction slice image. 4. The system of claim 3 wherein the difference image features are computed for those voxels that are associated with the cell including an average difference for the portion of the images containing the cell. 5. The system of claim 1 wherein the calculation of the selected features includes measuring an error between the cosine function and the center of mass trend. 6. The system of claim 5 wherein the selected features include the absolute and radius normalized maximum deviation, and root mean square error (RMSE) between the cosine function and the center of mass trend. 7. The system of claim 1 wherein the selected features are selected from the group consisting of histogram statistics, texture features and spatial frequency features. 8. The system of claim 1 wherein the object comprises a biological cell. 9. A method for detecting poor quality images in an optical tomography system comprising: operating the optical tomography system to acquire a set of pseudo-projection images of an object having a center of mass, where each of the set of pseudo-projection images is acquired at a different angle of view;transmitting the set of pseudo-projection images to a processor for reconstructing the pseudo-projection images into 3D reconstruction images; andoperating the processor to classify the 3D reconstruction images using selected features that characterize poor quality reconstructions wherein the selected features include difference image features that are formed by creating a difference image between a fixed focus image and a reconstruction slice image. 10. The method of claim 9 wherein the selected features describe streaking in the 3D reconstruction images. 11. The method of claim 9 wherein the difference image features are computed for those voxels that are associated with the cell including an average difference for the portion of the images containing the cell. 12. The method of claim 9 wherein the set of pseudo-projection images present a center of mass trend for the object and the selected features are calculated from a comparison of the center of mass trend with a cosine function fitting the trend in the center of mass. 13. The method of claim 12 wherein the calculation of the selected features includes measuring an error between the cosine function and the center of mass trend. 14. The method of claim 13 wherein the selected features include the absolute and radius normalized maximum deviation, and root mean square error (RMSE) between the cosine function and the center of mass trend. 15. The method of claim 9 wherein the selected features are selected from the group consisting of histogram statistics, texture features and spatial frequency features. 16. The method of claim 9 wherein the object comprises a biological cell. 17. A system for detecting poor quality images in an optical tomography system where a set of cells are suspended in an index of refraction matching gel contained in a rotating capillary tube, where pressure is applied to the gel to move the cells into the optical path of a microscope including an objective lens that is scanned through the capillary tube while an illumination source operates to illuminate cells passing through the field of view of the objective lens and where the optical tomography system includes an image sensor located to acquire pseudo-projection images transmitted from the objective lens, where each of the set of pseudo-projection images is acquired at a different angle of view, the system comprising: a reconstruction processor, coupled to receive a set of pseudo-projection images of each of the set of cells, each cell having a center of mass, where the reconstruction processor creates at least one 3D reconstruction image;a quality score classifier coupled to receive the at least one 3D reconstruction image, where the quality score classifier scores selected features in the at least one 3D reconstruction image;wherein the selected features describe streaking in reconstruction;wherein the set of pseudo-projection images present a center of mass trend for the cell and the selected features further include comparison features calculated from a comparison of the center of mass trend with a cosine fit curve;wherein the selected features include difference image features that are formed by creating a difference image between a fixed focus image and a reconstruction slice images; andwherein the difference image features are computed for voxels that are associated with the cell including an average difference for the portion of the images containing the cell. 18. The system of claim 17 wherein the calculation of the selected features includes selected features includes measuring an error between the cosine function and the center of mass trend. 19. The system of claim 18 wherein the selected features include the absolute and radius normalized maximum deviation, and root mean square error (RMSE) between the cosine function and the center of mass trend. 20. The system of claim 17 wherein the selected features are selected from the group consisting of histogram statistics, texture features and spatial frequency features.