Alternative Non-Destructive Approach for Estimating Morphometric Measurements of Chicken Eggs from Tomographic Images with Computer Vision

Abstract

The egg has natural barriers that prevent microbiological contamination and promote food safety. The use of non-destructive methods to obtain morphometric measurements of chicken eggs has the potential to replace traditional invasive techniques, offering greater efficiency and accuracy. This paper aims to demonstrate that estimates derived from non-invasive approaches, such as 3D computed tomography (CT) image analysis, can be comparable to conventional destructive methods. To achieve this goal, two widely recognized deep learning architectures, U-Net 3D and Fully Convolutional Networks (FCN) 3D, were modeled to segment and analyze 3D CT images of chicken eggs. A dataset of real CT images was created and labeled, allowing the extraction of important morphometric measurements, including height, width, shell thickness, and volume. The models achieved an accuracy of up to 98.69%, demonstrating their effectiveness compared to results from manual measurements. These findings highlight the potential of CT image analysis, combined with deep learning, as a non-invasive alternative in industrial and research settings. This approach not only minimizes the need for invasive procedures but also offers a scalable and reliable method for egg quality assessment.

Keywords: 3D image segmentation; computer tomographic images; deep learning; eggs quality; morphometric data extraction; poultry.

Figure 1

Measuring Equipment: (a) Egg measurements, length and width measured with a digital caliper (b) Eggshell thickness, measured with digital micrometer (c) Acquisition of images of chicken eggs by computed tomography.

Figure 2

Labeling of egg structures made in CVAT.

Figure 3

Pipeline of Estimated Morphometrics Measurements since the Inputs (Images) to Outputs (Estimated Masks and Morphometrics Measures) of the Models.

Figure 4

Approach to counting the number of voxels for estimating morphometric measurements.

Figure 5

Metrics of the segmentation stage: (a) Accuracy metrics of the training and validation sets of the U-Net and FCN models. (b) Loss metrics of the training and validation sets of the U-Net and FCN models.

Figure 6

Comparison between a slice (right) with low accuracy and a slice (left) with high accuracy.

Figure 7

Comparison of real tomographic image slices (column 1), corresponding labeled mask slices (column 2), probability outputs of the networks for each class (column 3–12), showing in yellow the most probable class and the final predictions for the U-Net and FCN networks (column 13–14).