3D characterization of walnut morphological traits using X-ray computed tomography

Toward robust and high-throughput detection of seed defects in X-ray images via deep learning

BACKGROUND

The detection of internal defects in seeds via non-destructive imaging techniques is a topic of high interest to optimize the quality of seed lots. In this context, X-ray imaging is especially suited. Recent studies have shown the feasibility of defect detection via deep learning models in 3D tomography images. We demonstrate the possibility of performing such deep learning-based analysis on 2D X-ray radiography for a faster yet robust method via the X-Robustifier pipeline proposed in this article.

RESULTS

2D X-ray images of both defective and defect-free seeds were acquired. A deep learning model based on state-of-the-art object detection neural networks is proposed. Specific data augmentation techniques are introduced to compensate for the low ratio of defects and increase the robustness to variation of the physical parameters of the X-ray imaging systems. The seed defects were accurately detected (F1-score >90%), surpassing human performance in computation time and error rates. The robustness of these models against the principal distortions commonly found in actual agro-industrial conditions is demonstrated, in particular, the robustness to physical noise, dimensionality reduction and the presence of seed coating.

CONCLUSION

This work provides a full pipeline to automatically detect common defects in seeds via 2D X-ray imaging. The method is illustrated on sugar beet and faba bean and could be efficiently extended to other species via the proposed generic X-ray data processing approach (X-Robustifier). Beyond a simple proof of feasibility, this constitutes important results toward the effective use in the routine of deep learning-based automatic detection of seed defects.

Characterizing the Internal Structure of Chinese Steamed Bread during Storage for Quality Evaluation Using X-ray Computer Tomography

Chinese steamed bread (CSB) is a traditional food of the Chinese nation, and the preservation of its quality and freshness during storage is very important for its industrial production. Therefore, it is necessary to study the storage characteristics of CSB. Non-destructive CT technology was utilized to characterize and visualize the microstructure of CSB during storage, and also to further study of quality changes. Two-dimensional and three-dimensional images of CSBs were obtained through X-ray scanning and 3D reconstruction. Morphological parameters of the microstructure of CSBs were acquired based on CT image using image processing methods. Additionally, commonly used physicochemical indexes (hardness, flexibility, moisture content) for the quality evaluation of CSBs were analyzed. Moreover, a correlation analysis was conducted based on the three-dimensional morphological parameters and physicochemical indexes of CSBs. The results showed that three-dimensional morphological parameters of CSBs were negatively correlated with moisture content (Pearson correlation coefficient range-0.86~-0.97) and positively correlated with hardness (Pearson correlation coefficient range-0.87~0.99). The results indicate the inspiring capability of CT in the storage quality evaluation of CSB, providing a potential analytical method for the detection of quality and freshness in the industrial production of CSB.

Research on the evolutionary history of the morphological structure of cotton seeds: a new perspective based on high-resolution micro-CT technology

Cotton (Gossypium hirsutum L.) seed morphological structure has a significant impact on the germination, growth and quality formation. However, the wide variation of cotton seed morphology makes it difficult to achieve quantitative analysis using traditional phenotype acquisition methods. In recent years, the application of micro-CT technology has made it possible to analyze the three-dimensional morphological structure of seeds, and has shown technical advantages in accurate identification of seed phenotypes. In this study, we reconstructed the seed morphological structure based on micro-CT technology, deep neural network Unet-3D model, and threshold segmentation methods, extracted 11 basics phenotypes traits, and constructed three new phenotype traits of seed coat specific surface area, seed coat thickness ratio and seed density ratio, using 102 cotton germplasm resources with clear year characteristics. Our results show that there is a significant positive correlation (P< 0.001) between the cotton seed size and that of the seed kernel and seed coat volume, with correlation coefficients ranging from 0.51 to 0.92, while the cavity volume has a lower correlation with other phenotype indicators (r<0.37, P< 0.001). Comparison of changes in Chinese self-bred varieties showed that seed volume, seed surface area, seed coat volume, cavity volume and seed coat thickness increased by 11.39%, 10.10%, 18.67%, 115.76% and 7.95%, respectively, while seed kernel volume, seed kernel surface area and seed fullness decreased by 7.01%, 0.72% and 16.25%. Combining with the results of cluster analysis, during the hundred-year cultivation history of cotton in China, it showed that the specific surface area of seed structure decreased by 1.27%, the relative thickness of seed coat increased by 8.70%, and the compactness of seed structure increased by 50.17%. Furthermore, the new indicators developed based on micro-CT technology can fully consider the three-dimensional morphological structure and cross-sectional characteristics among the indicators and reflect technical advantages. In this study, we constructed a microscopic phenotype research system for cotton seeds, revealing the morphological changes of cotton seeds with the year in China and providing a theoretical basis for the quantitative analysis and evaluation of seed morphology.

Nondestructive 3D phenotyping method of passion fruit based on X-ray micro-computed tomography and deep learning

Passion fruit is a tropical liana of the Passiflora family that is commonly planted throughout the world due to its abundance of nutrients and industrial value. Researchers are committed to exploring the relationship between phenotype and genotype to promote the improvement of passion fruit varieties. However, the traditional manual phenotyping methods have shortcomings in accuracy, objectivity, and measurement efficiency when obtaining large quantities of personal data on passion fruit, especially internal organization data. This study selected samples of passion fruit from three widely grown cultivars, which differed significantly in fruit shape, size, and other morphological traits. A Micro-CT system was developed to perform fully automated nondestructive imaging of the samples to obtain 3D models of passion fruit. A designed label generation method and segmentation method based on U-Net model were used to distinguish different tissues in the samples. Finally, fourteen traits, including fruit volume, surface area, length and width, sarcocarp volume, pericarp thickness, and traits of fruit type, were automatically calculated. The experimental results show that the segmentation accuracy of the deep learning model reaches more than 0.95. Compared with the manual measurements, the mean absolute percentage error of the fruit width and length measurements by the Micro-CT system was 1.94% and 2.89%, respectively, and the squares of the correlation coefficients were 0.96 and 0.93. It shows that the measurement accuracy of external traits of passion fruit is comparable to manual operations, and the measurement of internal traits is more reliable because of the nondestructive characteristics of our method. According to the statistical data of the whole samples, the Pearson analysis method was used, and the results indicated specific correlations among fourteen phenotypic traits of passion fruit. At the same time, the results of the principal component analysis illustrated that the comprehensive quality of passion fruit could be scored using this method, which will help to screen for high-quality passion fruit samples with large sizes and high sarcocarp content. The results of this study will firstly provide a nondestructive method for more accurate and efficient automatic acquisition of comprehensive phenotypic traits of passion fruit and have the potential to be extended to more fruit crops. The preliminary study of the correlation between the characteristics of passion fruit can also provide a particular reference value for molecular breeding and comprehensive quality evaluation.

Introducing Three-Dimensional Scanning for Phenotyping of Olive Fruits Based on an Extensive Germplasm Survey

Morphological characterization of olive (Olea europaea L.) varieties to detect desirable traits has been based on the training of expert panels and implementation of laborious multiyear measurements with limitations in accuracy and throughput of measurements. The present study compares two- and three-dimensional imaging systems for phenotyping a large dataset of 50 olive varieties maintained in the National Germplasm Depository of Greece, employing this technology for the first time in olive fruit and endocarps. The olive varieties employed for the present study exhibited high phenotypic variation, particularly for the endocarp shadow area, which ranged from 0.17−3.34 cm2 as evaluated via 2D and 0.32−2.59 cm2 as determined by 3D scanning. We found significant positive correlations (p < 0.001) between the two methods for eight quantitative morphological traits using the Pearson correlation coefficient. The highest correlation between the two methods was detected for the endocarp length (r = 1) and width (r = 1) followed by the fruit length (r = 0.9865), mucro length (r = 0.9631), fruit shadow area (r = 0.9573), fruit width (r = 0.9480), nipple length (r = 0.9441), and endocarp area (r = 0.9184). The present study unraveled novel morphological indicators of olive fruits and endocarps such as volume, total area, up- and down-skin area, and center of gravity using 3D scanning. The highest volume and area regarding both endocarp and fruit were observed for ‘Gaidourelia’. This methodology could be integrated into existing olive breeding programs, especially when the speed of scanning increases. Another potential future application could be assessing olive fruit quality on the trees or in the processing facilities.

A Computationally Efficient Reconstruction Algorithm for Circular Cone-Beam Computed Tomography Using Shallow Neural Networks

Circular cone-beam (CCB) Computed Tomography (CT) has become an integral part of industrial quality control, materials science and medical imaging. The need to acquire and process each scan in a short time naturally leads to trade-offs between speed and reconstruction quality, creating a need for fast reconstruction algorithms capable of creating accurate reconstructions from limited data. In this paper, we introduce the Neural Network Feldkamp-Davis-Kress (NN-FDK) algorithm. This algorithm adds a machine learning component to the FDK algorithm to improve its reconstruction accuracy while maintaining its computational efficiency. Moreover, the NN-FDK algorithm is designed such that it has low training data requirements and is fast to train. This ensures that the proposed algorithm can be used to improve image quality in high-throughput CT scanning settings, where FDK is currently used to keep pace with the acquisition speed using readily available computational resources. We compare the NN-FDK algorithm to two standard CT reconstruction algorithms and to two popular deep neural networks trained to remove reconstruction artifacts from the 2D slices of an FDK reconstruction. We show that the NN-FDK reconstruction algorithm is substantially faster in computing a reconstruction than all the tested alternative methods except for the standard FDK algorithm and we show it can compute accurate CCB CT reconstructions in cases of high noise, a low number of projection angles or large cone angles. Moreover, we show that the training time of an NN-FDK network is orders of magnitude lower than the considered deep neural networks, with only a slight reduction in reconstruction accuracy.

Genome-Wide Association Study Reveals Candidate Genes Involved in Fruit Trait Variation in Persian Walnut (Juglans regia L.)

Elucidating the genetic determinants of fruit quality traits in walnut is essential to breed new cultivars meeting the producers and consumers' needs. We conducted a genome-wide association study (GWAS) using multi-locus models in a panel of 170 accessions of Juglans regia from the INRAE walnut germplasm collection, previously genotyped using the Axiom^TM J. regia 700K SNP array. We phenotyped the panel for 25 fruit traits related to morphometrics, shape, volume, weight, ease of cracking, and nutritional composition. We found more than 60 marker-trait associations (MTAs), including a highly significant SNP associated with nut face diameter, nut volume and kernel volume on chromosome 14, and 5 additional associations were detected for walnut weight. We proposed several candidate genes involved in nut characteristics, such as a gene coding for a beta-galactosidase linked to several size-related traits and known to be involved in fruit development in other species. We also confirmed associations on chromosomes 5 and 11 with nut suture strength, recently reported by the University of California, Davis. Our results enhance knowledge of the genetic control of important agronomic traits related to fruit quality in walnut, and pave the way for the development of molecular markers for future assisted selection.