About

Dr. Cranos Williams is a Goodnight Distinguished Professor in Agricultural Analytics at NC State University. His role involves leading research in agricultural analytics, contributing to the advancement of knowledge in this field. As a faculty member, he is actively engaged in research activities, mentoring graduate students, and collaborating with other scholars to develop innovative solutions in agricultural data analysis and related areas. His work is recognized for its impact on agricultural sciences and analytics, and he holds a prominent position within the university's research community.

Research topics

• Artificial Intelligence
• Computer Science
• Engineering
• Computational biology
• Biology
• Machine Learning
• Biotechnology
• Genetics
• Botany
• Ecology
• Biochemistry
• Pulp and paper industry
• Mathematics

Selected publications

• Deployment and analysis of instance segmentation algorithm for in‐field yield estimation of sweet potatoes

  The Plant Phenome Journal · 2025-09-21 · 1 citations

  articleOpen access

  Abstract Shape estimation of sweet potato (SP) storage roots is inherently challenging due to their varied size and shape characteristics. Even measuring “simple” metrics, such as length and diameter, requires significant time investments either directly in‐field or afterward using automated graders. We present the results of a model that can perform grading and provide yield estimates directly in the field faster than manual measurements. Detectron2, a library consisting of deep‐learning object detection algorithms, was used to implement Mask region‐based convolutional neural network, an instance segmentation model. This model was deployed for in‐field grade estimation of SP roots and evaluated against an optical sorter. Roots from various clones, imaged with a cell phone during trials between 2019 and 2020, were used in the model's training and validation to fine‐tune a model to detect SP roots. Our results showed that the model (average precision = 74.1) could distinguish individual roots in environmental conditions, including variations in lighting and soil characteristics. Root mean square error (RMSE) for length, diameter, and weight, from the model compared to a commercial optical sorter, were 0.66 cm, 1.22 cm, and 74.73 g, respectively, while the RMSE of root counts per plot was 5.27 roots, with R = 0.8. This phenotyping strategy has the potential to enable rapid yield estimates in the field without the need for sophisticated and costly sorters and may be more readily deployed in environments with limited access to these resources or facilities.

  PublisherDOI

• Spatiotemporal dynamics of NF-κB/Dorsal inhibitor IκBα/Cactus in Drosophila blastoderm embryos

  iScience · 2025-06-09 · 3 citations

  articleOpen access

  embryos and highlights the need for similar quantitative studies in other biological systems.

  PublisherDOI

• Data from: Deployment and analysis of instance segmentation algorithm for in-field yield estimation of sweet potatoes

  Open MIND · 2025-10-26

  dataset

  Shape estimation of sweetpotato (SP) storage roots is inherently challenging due to their varied size and shape characteristics. Even measuring “simple” metrics, such as length and diameter, requires significant time investments either directly in-field or afterward using automated graders. We present the results of a model that can perform grading and provide yield estimates directly in the field faster than manual measurements. Detectron2, a library consisting of deep-learning object detection algorithms, was used to implement Mask R-CNN, an instance segmentation model. This model was deployed for in-field grade estimation of SP roots and evaluated against an optical sorter. Roots from various clones imaged with a cellphone during trials between 2019 and 2020, were used in the model’s training and validation to fine-tune a model to detect SP roots. Our results showed that the model (Average Precision = 74.1) could distinguish individual roots in environmental conditions, including variations in lighting and soil characteristics. Root mean square error (RMSE) for length, diameter, and weight, from the model compared to a commercial optical sorter, were 0.66 cm, 1.22 cm, and 74.73 g, respectively, while the RMSE of root counts per plot was 5.27 roots, with R^2 = 0.8. This phenotyping strategy has the potential to enable rapid yield estimates in the field without the need for sophisticated and costly sorters and may be more readily deployed in environments with limited access to these resources or facilities.

  DOI

• No more laborious stem counting: AI-powered computer vision enables identification and quantification of solid and hollow alfalfa stems at the pixel level

  Smart Agricultural Technology · 2025-08-05 · 1 citations

  articleOpen access

  • Novel thresholding of irregular polygons using YOLO, Otsu, and k-means learning • First-of-its-kind approach to quantify stem hollowness scores at the pixel level • Converting subjective to objective stem phenotyping with high accuracy • Automating stem hollowness phenotyping and data collection with high efficiency Traditional alfalfa stem phenotyping is labor-intensive and susceptible to bias from subjective ratings. Computer vision and machine learning present a promising solution for objectively assessing stem morphology. This study proposed an AI-driven image analysis to replace manual phenotyping methods with high efficiency and accuracy. We developed a novel pipeline that combines YOLOv8n with Otsu’s thresholding and K-means clustering to identify the medoids of internal and external polygons of the stem, thereby quantifying stem traits using pixel-based morphometric masks. The approach achieved an F1 score of 0.91 in detecting and classifying hollow or solid stems across plots and genotypes. Further analysis measured stem area and the proportions of stem tissue versus hollow regions, generating traits like hollowness score and percentage of hollowness. These stem-level metrics provide novel, objective, and quantitative phenotypic measurements, supporting ongoing chemical digestibility analyses and enabling real-time, image-based digestibility assessments through a field-deployable mobile application.

  PublisherDOI

• KeySDL: Sparse Dictionary Learning for Keystone Microbe Identification

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-08-11

  preprintOpen accessSenior authorCorresponding

  Identification of microbes with large impacts on their microbial communities, known as keystone microbes, is a topic of long-standing interest in microbiome research. However, many approaches to identify keystone microbes are limited by the inherent nonlinearity and state-dependence of microbial dynamics. Machine learning approaches have been applied to address these shortcomings but often require more data than is available for a given microbial system. We propose a keystone identification approach called KeySDL which reduces the amount of data required by incorporating assumptions about the type of microbial dynamics present in the experimental system. The data are modeled as originating from a Generalized Lotka-Volterra (GLV) model, an architecture commonly used to simulate microbial systems. The parameters of this model are then estimated using Sparse Dictionary Learning (SDL) Compared to existing methods, this approach allows accurate prediction of keystone microbes from small numbers of samples and provides an output interpretable as reconstructed system dynamics. We also propose a self-consistency score to help evaluate whether the assumption of GLV dynamics is reasonable for a given dataset, either through the application of KeySDL or other analysis tools validated using GLV simulation.

  PublisherDOI

• High‐throughput classification and quantification of skinning phenotype in sweet potatoes

  The Plant Phenome Journal · 2025-05-29 · 1 citations

  articleOpen access

  Abstract Sweet potatoes (SPs) ( Ipomoea batatas ) are a valued crop for their color, flavor, and nutrition. Harvesting is labor‐intensive, requiring hand‐picking to maintain skin quality. Mechanical harvesting often causes skin damage, known as “skinning,” where skin is cut, scraped, or torn, leading to lower quality during packing. To manage this, packers may conduct a costly “field‐switch” to reduce skinning in the production line. Currently, skinning levels are visually assessed by the packers and stakeholders. Field‐switches involve transitioning between multiple fields during harvest to meet specific customer orders (e.g., supermarkets, processing plants, or end users) that require high‐quality SPs. This process aims to minimize skinning and ensure the SPs meet the desired quality standards for those orders. This study introduces a computer vision (CV) pipeline to automate skinning assessment using a ResNet50‐based DeepLabV3+ semantic segmentation model. The CV system was trained to identify three classes: skinning, intact skin, and background. A machine vision camera, mounted above a conveyor belt, captured images throughout several full production days. The pipeline calculated the percentage of skinning () as the ratio between the predicted skinning area () and the total SP surface area (). Using this method, daily production trends and field‐switch decisions were studied with image data from six production days, chosen by our grower collaborator. Percent skinning ratings calculated from random subsamples of imaged SPs—where only a portion of the full image set was analyzed—showed no significant differences compared to those derived from complete image sets. This demonstrates that subsampling can reduce computational processing times by 90% while maintaining accuracy. When data were binned in 30‐min intervals, field‐switches occurred when there was approximately 1.5 PS. Across 2,417,907 SP instances, the model achieved a root mean square error of 0.55%, R of 0.84, 80.03% recall, 99.99% specificity, 77.44% F1 score, and 99.98% grading accuracy. This offers a promising improvement for automatic skinning detection on a commercial scale.

  PublisherOA PDFDOI

• Iron deficiency changes regulatory mechanisms governing sieve element cell differentiation

  Nature Communications · 2025-11-20

  articleOpen access

  Plant cell differentiation incorporates environmental cues over time to optimize overall growth. Iron deficiency influences development, such as root hair, cortical, and endodermal cell differentiation. However, the mechanisms by which iron deficiency regulates cell differentiation are not well characterized. Root sieve elements serve as an excellent model for cell differentiation since all cells, from undifferentiated to differentiated, are present in a distinct cell file. Here, we use semi-automated image analysis to show that iron deficiency delays sieve element differentiation, particularly enucleation and cell wall thickening, and consequently delays phloem sap unloading to roots. Using Dynamic Bayesian modeling we also characterize how iron deficiency changes the fundamental structure of the gene regulatory network associated with sieve element differentiation. We identify DOF1.5 as a positive regulator of sieve element enucleation and, consequently, of root sap translocation. These results clarify how abiotic stress can influence overall plant growth as a consequence of negatively influencing vascular differentiation. Plant growth and development respond to environmental cues. Here the authors infer gene regulatory networks controlling sieve element differentiation and show that under iron deficiency, enucleation and, consequently, phloem sap translocation are delayed.

  PublisherOA PDFDOI

• Knowledge and Opportunities for Managing Plant-Parasitic Nematodes Using Decision Intelligence

  PhytoFrontiers™ · 2025-04-09 · 1 citations

  articleOpen accessSenior author

  Plant-parasitic nematodes cause significant yield losses and economic damage to crops worldwide. Traditional management strategies are costly and environmentally unsustainable. This study introduces decision intelligence (DI) to enhance nematode management in sweetpotato cultivation. DI integrates decision analysis, artificial intelligence, and machine learning to support decision-making in complex environments. Through expert elicitations with researchers and extension agents, causal decision diagrams and causal decision models were developed for nematode management. These models were refined and tested at a field day event, revealing critical actions, intermediates, and external factors that influence outcomes, with profitability identified as the key outcome. Key intermediates included nematode soil populations in subsequent years, influenced by actions such as fumigation, planting a susceptible target crop (sweetpotatoes), or planting a non-susceptible rotational crop. External factors included soil temperature, moisture, and the costs associated with different management choices. Development of the causal decision model highlighted gaps in existing studies, particularly the need for data linking pesticide soil concentrations to nematode reduction under varying soil conditions. This research demonstrates the potential benefits of DI in plant pathology, including the ability to (i) engage stakeholders early in the modeling process, (ii) visualize complex domain-specific problems for interdisciplinary teams, and (iii) provide immediate value in understanding decision-making processes to both stakeholders and researchers. Overall, DI offers a promising approach to enhancing agricultural decision-making and has potential applications in plant pathology. [Formula: see text] Copyright © 2025 The Author(s). This is an open access article distributed under the CC BY-NC-ND 4.0 International license .

  PublisherDOI

• In-season yield forecasting using multitemporal remote sensing environmental observations and machine learning: Applications for sweetpotato in North Carolina, USA

  European Journal of Agronomy · 2025-09-10 · 1 citations

  article
  PublisherDOI

• Machine Learning for Predicting Sweetpotato Growth and Quality: Integrating In-Season Root Imaging with Environmental and Agronomic Insights

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

Recent grants

• INSPIRE: Dynamic Regulatory Modeling of the Iron Deficiency Response in Arabidopsis thaliana

  NSF · $1000k · 2012–2019

Frequent coauthors

• Jack Wang

  Louisiana State University Health Sciences Center New Orleans

  26 shared

• Vincent L. Chiang

  Northeast Forestry University

  26 shared

• Ronald R. Sederoff

  16 shared

• Quanzi Li

  Chinese Academy of Forestry

  15 shared

• Wolfgang Busch

  Salk Institute for Biological Studies

  12 shared

• Gabriel Krouk

  Institut des Sciences des Plantes de Montpellier

  12 shared

• W.E. Alexander

  12 shared

• Rosangela Sozzani

  North Carolina State University

  11 shared

Labs

• EnBiSys LabPI

  Using multiscale modeling to translate lab findings to field results through interdisciplinary plant science and agriculture data analytics.

Education

• Ph.D., Electrical and Computer Engineering

  North Carolina State University

  2008

• M.S., Electrical and Computer Engineering

  North Carolina State University

  2002

• B.S., Electrical and Computer Engineering

  North Carolina Agricultural and Technical State University

  2001