Unbiased sampling and thresholding in scientific research stereology

This video goes over methods for unbiased sampling taken from the field of stereology. It also includes methods for determining thresholds for images without user bias. The use of Photoshop and ImageJ makes up part of the discussion, as well as choosing the appropriate number of sections from a biological sample, enough animals, and an appropriate number of fields from microscope images. Setting a threshold is important for 3D reconstructions and for quantitation, image analysis and morphometry. A method for fluorescent images includes choosing non-specific stained samples to find the brightest of these, and then choosing a threshold level based upon those values. For grayscale and EM images, a method for color match in Photoshop allows for matching histograms so all images are at the same brightness and contrast level. For all images a high pass filter can be set based upon the length of significant objects in a scientific image, the filter is applied, and then a consistent threshold level can be found for application to all relevant images. A visual threshold can be set, though this is subjective. However, it can be made more objective by practicing setting a threshold against a known reference image. A highly cited article in "Science" proves its merit. The presentation is given by Jerry Sedgewick, a former director for a microscopy core lab at the University of Minnesota. His accomplishments include publication as co-author for quantization methods in Science, PNAS, Circulation, and other journals. He has written two books and several book chapters on Photoshop, digital imaging, and image ethics. Mr. Sedgewick has given several workshops on scientific imaging, including at Woods Hole, MA. Instrumentation built by Mr Sedgewick include a multi-photon microscope and an imaging system developed for a proteomics device (Isoplexis) for which he was awarded a patent. Forensics work for a law agency on doctored images occupied a period of time in his career.