A recent study examines a new Internet of Things (IoT) framework using visible and near-infrared spectroscopy technology to accurately delineate between male and female pigeon eggs.
By using visible and near-infrared (NIR) spectroscopy technology with advanced neural network modeling, the food industry can effectively distinguish male and female pigeon eggs, improving breeding efficiency and commercial operations, according to a recent study published in Microchemical Journal (1).
Pigeon keeping is the art and science of breeding pigeons for domestic purposes, whether it is for meat or sport (2). Although fossil remains have taught us that these pigeons evolved in south Asia, pigeons are now bred all over the world (2). In the pigeon breeding industry, the ability to distinguish between male and female pigeons early in their development is important for economic reasons. Traditionally, breeders have had to wait until the pigeons hatched to determine their sex, which often led to delays in managing the freshness of pigeon eggs and disrupted sales plans (1). However, with the advent of new technology, researchers in this field are seeking to explore ways how they can determine pigeon sex before hatching to save time.
City Ledge. Aerial View of NYC Pigeon on Altitude Ledge in the Urban Background | Image Credit: © Popelniushka - stock.adobe.com
Lead author Huazhou Chen and his team at Guilin University of Technology explored this issue in their recent study. Their study introduced a novel Internet of Things (IoT) framework that could help the livestock industry distinguish between male and female pigeon eggs before hatching. Their method uses spectroscopic techniques to great effect. By using visible and near-infrared (Vis-NIR) spectroscopy technology combined with advanced neural network modeling, the researchers tested whether this method could timely and accurately determine the sex of pigeon eggs (1).
For their study, the research team constructed an IoT framework specifically designed to model and discriminate batches of pigeon eggs using data from Vis-NIR spectroscopy technology (1). This approach involved using multi-locational Vis-NIR sensors to monitor the spectral detection data of pigeon eggs (1). The data is then immediately transmitted to a cloud network unit within the IoT framework. At the heart of this system is a random weight neural network (RWNN) architecture, which serves as the intelligent computing module responsible for model training and optimization.
One of the standout features of the RWNN architecture is its adaptive learning strategy. This strategy is designed to fine-tune the network linkage weights and relevant hyperparameters, ensuring that the model remains highly accurate in processing the constant inflow of Vis-NIR big data (1). Additionally, the researchers embedded a partial least squares discriminant analysis (PLS-DA) method within the Softmax unit of the neural network (1). By integrating PLS-DA within the Softmax unit, the researchers were able to optimize data processing with spectral properties, which in turn helped improve accuracy of sex discrimination of pigeon eggs (1).
To validate their IoT-based RWNN framework, Chen and his team conducted extensive experiments using the Vis-NIR data of 309 pigeon eggs. The eggs were monitored on the 5th, 6th, 7th, and 8th days of hatching, and the RWNN model was tested with various network structures (1). The results indicated that the adaptive RWNN architecture was able to achieve a high prediction accuracy in distinguishing between male and female pigeon eggs at these early stages of development (1).
As a result, Chen’s team demonstrated in their study the utility of their IoT-based RWNN framework. Apart from determining sex differences of pigeon eggs, the method proposed by Chen and colleagues can also be used in other livestock applications where sex determination is vital. The integration of Vis-NIR spectroscopy with advanced neural network modeling offers a scalable and efficient solution for processing large volumes of spectral data, paving the way for broader applications in the agricultural sector (1).
The livestock industry continues to explore how technology can help improve their standard business practices. As the IoT-based Vis-NIR technology in conjunction with the adaptive RWNN architecture demonstrates in this study, it can be used to continually improve the efficiency and profitability of breeding operations.
(1) Cai, K.; Fang, Q.; Lin, Q.; et al. Fast Discrimination of Female and Male Pigeon Eggs Using Internet of Things in Combined with Vis-NIR Spectroscopy and Chemometrics. Microchem. J. 2024, 203, 110883. DOI: 10.1016/j.microc.2024.110883
(2) Gilbert, M. T. P.; Shapiro, M. D. Pigeons: Domestication. In: Encyclopedia of Global Archaeology. Smith, C, Ed. Springer: New York, 2014, pp. 5944–5948. DOI: 10.1007/978-1-4419-0465-2_2214
Best of the Week: AI, Rapid Food Analysis, Agriculture Analysis, and Soil Property Prediction
September 6th 2024Top articles published this week include a review article on the latest research in agriculture analysis, a peer-reviewed article on near-infrared (NIR) spectroscopy, and an interview about using fluorescence spectroscopy in cheese ripening.
AI-Powered Spectroscopy Faces Hurdles in Rapid Food Analysis
September 4th 2024A recent study reveals on the challenges and limitations of AI-driven spectroscopy methods for rapid food analysis. Despite the promise of these technologies, issues like small sample sizes, misuse of advanced modeling techniques, and validation problems hinder their effectiveness. The authors suggest guidelines for improving accuracy and reliability in both research and industrial settings.
Non-Linear Memory-Based Learning Advances Soil Property Prediction Using vis-NIR Spectral Data
September 3rd 2024Researchers from Zhejiang University have developed a new non-linear memory-based learning (N-MBL) model that enhances the prediction accuracy of soil properties using visible near-infrared (vis-NIR) spectroscopy. By comparing N-MBL with traditional machine learning and local modeling methods, the study reveals its superior performance, particularly in predicting soil organic matter and total nitrogen.
Measuring Muscle Oxidation in Athletes with Near-Infrared Spectroscopy
September 2nd 2024A recent article authored by scientists from the Institute of Sport and Preventive Medicine, part of the University of Saarland (Saarbrücken, Germany), discusses their investigation of the absolute and relative test-retest reliability of the Moxy Monitor, as well as their investigations into side differences of oxygen saturation at the vastus lateralis muscle of both legs in male cyclists.