자기공명영상을 활용한 무정란 조기 판별 기술 개발

다국어 초록 (Multilingual Abstract)

The objectives of this study were to develop an image processing algorithm that can observe image characteristics that change with incubation time, to develop shape indices that can distinguish between fertile and infertile eggs using image characteristics, to investigate whether normal hatching can be determined using these indices, and to develop a technology capable of detecting infertile eggs. MRI acquisition was performed using an industrial MRI system and 40 fertilized and 10 infertile eggs. Eggs were placed in an incubator at 38°C for 3 days, and MR images were obtained 8 times at baseline (before samples were placed in the incubator) and at 22, 32, 40, 48, 56, 64, and 72 hours after placing samples in the incubator. An image processing algorithm was developed to analyze centroid, circularity, and the major-to-minor axis ratios to quantify yolk shape changes during incubation from extracted images. Average image information of three central slices was used to calculate shape indices useful for differentiating fertile and infertile eggs. The shape indices yolk circularity and major-to-minor axis ratio (determined after incubation for 56 hr) achieved discrimination rates of 98.4% and 82.5%, respectively. In summary, using shape indices would enable fertile and infertile eggs to be differentiated within 3 days.