International Journal of Systematic Innovation

Evaluation of convolutional neural network models' performance for estimating mango crop yield

Neethi M V, P. Raviraj — Wed, 19 Feb 2025 00:00:00 +0800

In agriculture, crop yield estimation is essential; producers, industrialists, and consumers all benefit from knowing the early yield. Mango manual counting typically involves the utilization of human labor. Experts visually examine each sample to complete the process, which is time-consuming, very difficult and has little precision. For commercial mango production to produce high-quality fruits from the orchard to the consumer, a quick, non-destructive, and accurate variety classification is required. Because of its effectiveness in computer vision, a convolutional neural network—one of the deep learning techniques—was chosen for this investigation. For yield prediction, a total of eight popular mango cultivars were utilized. A comparison with previously trained models was used to assess the suggested model.The performance of the classifiers was evaluated using evaluation metrics such as accuracy, loss, Roc-AUC score, precision, recall, F1-score, sensitivity, specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), and Cohen-Kappa performance measures. In terms of performance evaluation criteria, it was discovered that the proposed approach performed better than the pre-trained models. It was discovered that the suggested model produced 98.85% accuracy in the test set, which had 800 images. This outcome has demonstrated the tangible applicability of the proposed methodology for mango crop estimation.

Technology innovation of dryer machine based on sustainability automation systems to increase agel fiber production in handicraft SME

Wed, 19 Feb 2025 00:00:00 +0800

In Indonesia, natural fibers are extensively utilized as essential raw materials for various human needs. These natural fibers find significant application in handicrafts, particularly in small, and medium enterprises (SME) within the handicraft industry. Agel fiber is obtained from drying gebang leaves, is a prominent natural fiber, plays a vital role in the local economy of the Kulon Progo region in Yogyakarta, Indonesia. Currently, the production process of raw materials to become fiber relies on conventional methods, primarily sun-drying, which often gives rise to numerous challenges such as temperature fluctuations and weather dependencies. Moreover, the physical posture adopted by workers during the drying process is ergonomically unfavorable, as they must bend over repeatedly to turn the Agel Fibers being dried under the sun. Recognizing these issues,ure it becomes imperative to develop a sustainable dryer machine equipped with advanced technology that enhances productivity while prioritizing employee ergonomics. This study employs the research and development (R&D) method, which encompasses analysis, design, development, implementation, and evaluation stages. The primary objective of this research is to design, fabricate, and test a dryer machine utilizing a sustainability automation system integrated with Internet of Things (IoT). The outcome of this research is a dryer machine that can effectively dry Agel leaves within a significantly reduced timeframe of 2-4 hours, with a maximum capacity of 10 kg per cycle. This achievement surpasses the conventional method, which typically takes 5-6 days to produce 10 kg of dried Agel Fibers.

Brightness Augmentation Implementation to Evaluate Perfor-mance Classification of Face Masked Base on CNN Model

Rianto, Desty Mustika Ramadhan, Husni Mubarok — Wed, 19 Feb 2025 00:00:00 +0800

The use of Deep Learning methods with CNN models is starting to be applied such as in Facial Expression Recognition. However, with the pandemic situation in recent years, there are still some people who wear masks for work purposes or because they are sick so that their faces are not fully visible. This can affect social interactions, especially in the mouth area which is very informative. Therefore, this research aims to provide a better understanding of masked facial expression recognition with the application of CNN models, namely with VGG16 and MobileNet architectures. In addition, this research will also explore the use of data augmentation methods, such as geometric augmentation and brightness augmentation, to see their effect on the classification accuracy of masked facial expressions. The results show that the use of VGG16 architecture with cross-validation method (VGG16-FLCV) provides better performance than MobileNet-FLCV architecture in recognizing and classifying masked facial expressions. The application of data augmentation methods, such as geometric augmentation and brightness augmentation, has helped to improve the performance of CNN models. Experimental results show that on the VGG16-FLCV architecture, the brightness range (1.00, 1.25) provides the best accuracy with a training accuracy of 81.73% and a validation accuracy of 70.71%. In addition, this study found that the optimal use of brightness range on the VGG16-FLCV architecture is in the darkness category with ranges (0.25, 0.50), (0.50, 0.75), and (0.75, 1.00), as well as in the brightness category with ranges (1.00, 1.25). This study found that the MobileNet-FLCV architecture with a brightness range of (0.25, 0.50), (0.50, 0.75), (0.75, 1.00), (1.00, 0.25), and (1.25, 1.50) can be used as an alternative brightness range that can still be applied without experiencing a significant decrease in accuracy.

A systematic meta-analysis on the role of artificial intelligence and machine learning in detection of gynaecological disorders

Jyoti Nandalwar, Pradip Jawandhiya — Wed, 19 Feb 2025 00:00:00 +0800

Gynaecological disorder is a serious health issue that affects women's health globally. The use of Artificial Intelligence (AI) or Machine learning (ML) techniques has gained the attention of researchers for the detection and diagnosis of gynaecological disorders such as cancer. This paper aims to provide insight into the role of AI in gynaecological disorder diagnosis. This paper also provides a systematic analysis of several AI/ML approaches that are being employed. The paper investigates how ML algorithms can extract characteristics from MRI images and how to use ML to extract and recognize the features from medical images such as MRI, ultrasound, CT-scans, etc. for early detection of gynaecological tumors and provide more personalized risk assessment. However, it is observed that there will be a significant impact of the advancement of AI/ML on medical technology in the future. Therefore, this paper presents a significant contribution to future medical applications and innovations.

Systematic modernization of fish smoking method with the implementation of smoked fish machine based on Internet of Things technology

Rizal Justian Setiawan, Khakam Ma'ruf; Darmono, Nur Azizah; Nur Evirda Khosyiati — Wed, 19 Feb 2025 00:00:00 +0800

The traditional process of smoking fish, which is widely used in coastal regions, poses significant challenges due to its labor-intensive nature, the need for constant supervision, and the difficulty in maintaining stable temperatures. These issues often result in inefficiencies, inconsistent product quality, and potential safety hazards. Given the importance of the smoked fish industry in sustaining local economies in the coastal area of Indonesia, there is a critical need for more advanced, reliable, and efficient methods of fish smoking. This study addresses these challenges by developing an Internet of Things (IoT)-integrated monitoring and control system for the smoked fish machine. This study was conducted to develop a monitoring and control system for machines which includes turning on/off machine components, temperature monitoring, and blower RPM control. The results of the study showed that the implementation of IoT can activate machine components such as a blower, a servo motor, and a light. Moreover, IoT can monitor the machine temperature from a smartphone in real-time by integrating with a temperature sensor. The temperature difference between the sensor and the analog thermometer was found to be 0.1 - 0.5^oC, it was proving that the temperature on the IoT system was not much different from the analog thermometer. Furthermore, the blower RPM control results showed that the system could maintain the temperature in the optimal range (75-90^oC) for smoking the fish, with a maximum deviation of 1°C, and can adjust the Blower RPM according to the user's wishes through control from the IoT system. In general, the use of IoT that has been developed can make it easier for users to operate the machine.

Knowledge management capability and innovation ambidexterity: The role of intellectual capital and intangible competitive advantage

Nirusa Sirivariskul — Wed, 19 Feb 2025 00:00:00 +0800

This study examines the impact of intellectual capital enabled knowledge capability management on innovation ambidexterity and the moderating role of intangible resources advantage through the lens of open innovation. The sample consists of 105 companies in the Thai food industry. To enhance understanding of how companies can achieve success in innovation, an interesting result of the study is that it helps companies explore new knowledge and leverage existing or new knowledge to innovate more continuously. Based on the presented findings, intangible resources advantage plays a positive moderating role in this relationship: knowledge capability management and intangible resources advantage work together to foster innovation ambidexterity.

Custom hardware design for peripheral artery disease detection: Field-programmable gate arrays and application-specific integrated circuits

Pravalika Nazarkar — Wed, 19 Feb 2025 00:00:00 +0800

Atherosclerotic disorders such as peripheral artery disease (PAD) has a major negative influence on patient outcomes. Inadequate treatment and a poor detection rate can result in cardiovascular problems and limb loss. There is great promise for improving the detection and treatment of PAD and other medical disorders with machine learning (ML) and artificial intelligence (AI) techniques. Field-Programmable Gate Arrays (FPGAs) and Application-Specific Integrated Circuits (ASICs) are used to implement the fundamental ideas of AI and ML, as this study highlights. We look at how these technologies are used for PAD, highlighting how they may be used to better pick drugs, improve patient care, and improve disease phenotype. Providing accurate and effective solutions for difficult medical problems, the fusion of AI and ML with FPGA and ASIC technology represents a significant breakthrough in medical analytics.

Optimized cross-corpus speech emotion recognition framework based on Normalized 1D Convolutional Neural Network

Nishant Barsainyan, Dileep Kumar Singh — Wed, 19 Feb 2025 00:00:00 +0800

Human-computer interaction (HCI) improved via voice detection of emotions. Speech Emotion Recognition (SER) software typically detects the appearance of various feelings in the speaker. However, there are significant challenges in combining information from multidisciplinary domains, notably speech-emotion recognition and applied psychology. Some researchers have used handcrafted attributes to categorize emotions and obtained high classification accuracy. However, these attributes reduce the categorization accuracy for multi-lingual environments. Deep learning algorithms have been utilized to autonomously retrieve the local representation from supplied speech data. The given strategies can't extract the most valuable characteristics from challenging speech inputs. To address this constraint, we propose an innovative SER framework that employs data augmentation approaches before generating relevant feature sets from each utterance and selecting the most discriminative optimum features. And the chosen feature vector is sent into the Normalized 1D CNN for emotion recognition using multi-lingual databases. This study evaluates the effectiveness of an XGB classifier for multi-lingual emotion recognition by testing its performance on data from a corpus trained on a different corpus. The testing outcomes displayed that our proposed SER architecture functioned better than existing SER approaches.

Predictive analytics: Unveiling the potential of machine learning and deep learning

kavitha P, Shakkeera L — Wed, 19 Feb 2025 00:00:00 +0800

Machine and deep learning methods have gained significant traction in the healthcare industry, particularly for the prediction of cardiac diseases. The increasing prevalence of heart-related diseases underscores the need for proactive and accurate health care interventions. Machine learning is a data-driven approach for actively recognizing and addressing cardiovascular risks. To achieve this, researchers have utilized a range of classification techniques, such as Support Vector Machines, Random Forests, and Naive Bayes, to disentangle the intricate aspects of heart disease prediction. Additionally, the Stacking Ensemble Learning Technique was used to further enhance prediction accuracy. However, the ensemble approach has certain limitations. Therefore, confusion matrices are utilized for thorough evaluation and validation, offering better classifier performance. As research advances, prediction models aim to achieve higher accuracy and generalizability. Insights from confusion matrices can help researchers to make more robust and dependable predictions. The implications of this research extend beyond academia and will benefit clinicians, patients, and healthcare systems. In conclusion, the confluence of machine learning, deep learning, and healthcare heralds a new era of precision medicine in which data-driven insights empower stakeholders to tackle formidable challenges with unparalleled effectiveness.

MNETGIDD: A heuristic-oriented segmentation and deep learning multi-disease detection model for gastrointestinal tracts

Bamini A — Wed, 19 Feb 2025 00:00:00 +0800

Malignant growth of the gastrointestinal tract is among the leading causes of death worldwide. Research indicates that almost 40% of people worldwide suffer from long-term digestive issues. According to a study published in the United European Gastroenterology Journal, the occurrence of a digestive disorders has increased since 2000. Digestive disorders continue to be a major cause of death even with a slight decline. The WHO MORTALITY DATABASE reported huge death rates in every year due to the GI Diseases. From that report, the need of an accurate detection of GI Tract malignant in low cost and error prone labor must be developed. This work introduces MNET Gastro Intestinal (GI) Disease Detection (MNETGIDD), which is a complete identification model for multi-gastrointestinal disease discovery from clinical images. MNETGIDD model Using Gastrolab dataset with endoscopic images and it works as pipelines that are pre-processed, segmented and identify the affected areas. This proposed approach aims to enhance image quality, facilitate accurate segmentation and classification, the entire process through a pipeline process, initially preprocessing with techniques such as text removal, illumination enhancement, and fuzzy histogram equalization. During segmentation, Otsu Segmentation based on Krill-Herd Optimization is used to identify the affected area. The MNETGIDD model incorporates the MobileNetV2 architecture, designed for light weight classification model working under resource-constrained environments. According to the tests, the MNETGIDD model exhibits high sensitivity and specificity, outperforming human experts in many cases. In terms of accuracy, the model achieved 96.349%, a precision 96.25 %, and a recall of 97.08%. This deep learning system has the potential to revolutionize gastrointestinal disease diagnostics and screening by automating key steps and improving patient outcomes.