| abstract |
A method of determining the reliability of a cytological screening test including the steps of viewing at least part of a cytological specimen, creating an image of the view, producing a digital representation of the image, ranking objects in the digital representation based on the degree to which each respective object has characteristics more likely found in a typical premalignant or malignant cell than in a typical benign cell, and displaying for review by an operator the images of at least a minimum number of the highest ranked objects such that objects of a predetermined cell type are displayed in the absence of premalignant or malignant cells, wherein the detection by the operator of the presence in the display of at least one cell of the predetermined cell type tends to indicate that the test is reliable. |