Abstract: A method for detecting malignancy-associated changes. A sample of cells is obtained and stained to identify the nuclear DNA material. The sample is imaged with a digital microscope. Objects of interest are identified in the sample of cells based on the intensity of the pixels that comprise the object versus the average intensity of all pixels in the slide image. An exact edge is located for each object and variations in the illumination intensity of the microscope are compensated for. A computer system calculates feature values for each object and, based on the value of the features, a determination is made whether the cell exhibits malignancy-associated changes or not.

Abstract: An imaging system for white light and fluorescence endoscopy that includes an automatic gain control circuit 30 that adjusts the brightness of an image produced based on distribution of pixel intensities in one or more video frames. The magnitude of the image signals produced by a pair of high sensitivity imaging devices such as intensified CCD transducers are compared to a number of reference thresholds. A time-over-threshold counter (112) determines the number of pixels in the image signals having magnitudes greater than or less than the reference thresholds. The distribution of pixel intensities is supplied to a decision tree algorithm (116) that determines whether the gain of the-intensified CCD transducers (44a, 44b) used to produced the autofluorescence images or the intensity of the excitation light produced by a light source (36) should be increased or decreased.

Abstract: An imaging system for white light and fluorescence endoscopy that includes an automatic gain control circuit 30 that adjusts the brightness of an image produced based on distribution of pixel intensities in one or more video frames. The magnitude of the image signals produced by a pair of high sensitivity imaging devices such as intensified CCD transducers are compared to a number of reference thresholds. A time-over-threshold counter (112) determines the number of pixels in the image signals having magnitudes greater than or less than the reference thresholds. The distribution of pixel intensities is supplied to a decision tree algorithm (116) that determines whether the gain of the-intensified CCD transducers (44a, 44b) used to produced the autofluorescence images or the intensity of the excitation light produced by a light source (36) should be increased or decreased.

Abstract: A method for detecting malignancy-associated changes. A sample of cells is obtained and stained to identify the nuclear DNA material. The sample is imaged with a digital microscope. Objects of interest are identified in the sample of cells based on the intensity of the pixels that comprise the object versus the average intensity of all pixels in the slide image. An exact edge is located for each object and variations in the illumination intensity of the microscope are compensated for. A computer system calculates feature values for each object and, based on the value of the features, a determination is made whether the cell exhibits malignancy-associated changes or not.

Abstract: Apparatus for imaging diseases in tissue comprising a light source for generating excitation light that includes wavelengths capable of generating characteristic autofluorescence for abnormal and normal tissue. A fiber-optic illuminating light guide is used to illuminate tissue with light that includes at least the excitation light thereby exciting the tissue to emit the characteristic autofluorescence. An imaging bundle collects emitted autofluorescence light from the tissue. The autofluorescence light is filtered into spectral bands in which the autofluorescence intensity for abnormal tissue is substantially different from normal tissue and the autofluorescence intensity for abnormal tissue is substantially similar to normal tissue. An optical system is used to intercept the filtered autofluorescence light to acquire at least two filtered emitted autofluorescence images of the tissue.

Abstract: A system for detecting cancerous or precancerous lesions directs light produced from a mercury arc lamp into an illumination guide of an endoscope. Autofluorescence light produced by the tissue under examination is divided into red and green spectral bands by a dichroic mirror. Light in the red and green spectral band is applied to a pair of image intensified CCD cameras. The output of the camera that receives light in the red spectral band is coupled to a red video input of a color video monitor. Light produced by the camera that receives light in the green spectral band is coupled to the blue and green video inputs of the video monitor. The system produces a false color display, whereby healthy tissue appears cyan in color and cancerous or precancerous lesions appear reddish in color. The image displayed allows the operator to see the lesions within the context of the underlying tissue structures.