EARLY PREDICTION OF PANCREATIC ABNORMALITIES USING GAIN METRIC

Pancreatic cancer and associated disorders are usually detected at their later stages, since their
early signs and symptoms are minimalor even unnoticeable. This paper proposed a new methodology for early detection based on Digital Image Processing concept. The proposed system follows a structured workflow comprising image acquisition, preprocessing in terms of contrast
enhancement and thresholding, and extraction of the Region of Interest. Accordingly, with the extracted values of GAIN, an image is classified as either normal or abnormal in pancreas, with the support of simple machine learning and statistical analysis. GAIN refers to the variance in intensity between normal and affected pancreatic images. By highlighting subtle pixel level variations, the
approach enhances detection of pancreatic abnormalities that could easily be missed through visual
observation. this technique is cost- effective , Non invasive in nature and can be integrated into any
imaging modality, thus supporting clinical practice at every level, from central hubs to peripheral
and remote diagnostics. Such a system will be of great help to healthcare professionals by
providing speedy and accurate diagnoses, which may lead to better outcomes in pancreatic care.