Microscope examination of Gram stained clinical specimens is used for aiding diagnosis of patients with infectious diseases. In high volumen pathology laboratories, this manual microscopy examination is considered time consuming and labour intensive. To that end, we propose a novel Gram stain image dataset, here called the UQ-SNP Gramstain Cell Detection (UQSNP_GCellDetect) dataset.
We use the leave-one-out cross validation Ten-fold cross validations.
A Multiple Covariance Approach for Cell Detection of Gram-Stained Smears ImagesMatthew Crossman, Arnold Wiliem, Anthony Jennings and Brian C. LovellInternational Conference on Acoustics, Speech and Signal Processing (ICASSP), 2015.