Contour Edge Mask Filtered Residual Network based Egg Classification

In agricultural and food safety, classifying eggs is a crucial activity that requires great precision in distinguishing various kinds of egg. Traditional approaches can be expensive and error-prone since they frequently rely on manual examination or simple image processing techniques. In order to improve classification performance, this research proposes Contour Edge Mask Filtered ResNet (CEM-ResNet), which combines advanced deep learning techniques with edge-detection methods. The Egg Dataset containing 2000 egg images that was used for implementation. The suggested CEM-ResNet model uses contour edge masking to eliminate extraneous background noise and highlights key structural information while utilizing ResNet architecture to capture essential details of eggs. The suggested CEM-ResNet model uses contour edge masking to eliminate extraneous background noise and highlights key structural information while utilizing ResNet architecture to capture essential details of eggs. Initially, the CEM-ResNet model organizes the egg into two groups as Chicken egg and Duck egg. The organized eggs are performed with labeling and then subjected to data augmentation to end up with 42,000 images. The augmented egg images are converted to grayscale images. The grayscale images are processed with edge detection model to finalize the contour filtered egg images and ResNet model. The contour filtered egg images are refined to create the edges and then the mask was created to generate Contour Edge Mask Filtered (CEM) egg images. The ResNet model was refined to process the CEM egg images. With a high accuracy of 99.64%, the implementation reveals that the proposed CEM-ResNet model outperforms towards the classification task.