Corn is an agricultural product that has been consumed either as food or alternative energy sources. To satisfy these needs, it is important to certainly obtain the best seed quality. The corn seeds are obtained from planted corn at particular environment. The seed quality might be vary due to soil and weather condition. Corn experts normally observe and determine corn seed quality based on their visual assessment. Corn kernels should be removed from the corn cob to obtain the seeds. The assessment can be difficult if the seeds are collected from various regions. Manual assessment is also influenced by subjectivity of the corn experts. Imaging based algorithm is proposed to assess corn seed quality. Region of interest selection is the initial step of quality assessment. In the selection, the algorithm have to correctly define corn seed location and the boundary box of the seeds. Automatic Region-of-Interest (ROI) selection is developed to provide images containing a single corn. Therefore, total number of ROI images will be similar with total number of seed in the image. The paper presents an algorithm for automatically selecting ROI of corn seed. Based on the evaluation results, automatic ROI selection has been proven with most of segmentation accuracies are higher than 90%. By applying this algorithm, the corn quality can then be analysed in every single seed.