Applying Artificial Intelligence Research Articles

BT08 Mapping UK frameworks for ethical artificial intelligence applied to dermatology

Abstract There is significant enthusiasm to apply artificial intelligence (AI) in dermatology with an emphasis on image-based analyses for risk assessment of suspicious skin lesions. However, the rapid development and deployment of such AI tools raises concerns that ethical issues arising from their widespread, long-term use may not have been adequately considered during design or implementation. Dermatology, in particular, has been identified as lacking sufficient bioethics research that could standardize the approach to regulating AI-assisted applications. Despite significant efforts to develop robust guidelines for ‘ethical AI’ across private, national and international organizations, debate continues surrounding which ethical principles should be included. The 2022 National AI Strategy positions the UK as leading global efforts toward achieving safe and ethical deployment of AI. We sought to systematically review and analyse policy documents to map UK principles for promoting ethical AI as applied to dermatology. Our review identified five UK policy documents relevant to ethical AI published between 2016 and 2022. All included frameworks highlighted the importance of upholding transparency and fairness along the entire lifecycle of AI development and decision-making. The five frameworks also incorporated autonomy through reference to maintaining informed consent, data privacy and human contestability of AI decisions. The principles of accountability and nonmaleficence were each identified in four policy documents, emphasizing the needs for clear definitions of legal and organizational responsibility and for thorough testing of AI tools to avoid unintended/unwanted outcomes. The principles of beneficence (designing AI to promote good on both individual and societal levels) and sustainability (continually monitoring and adjusting AI tools for longevity and efficacy) were each included in two policy documents. Current UK policy frameworks for ethical AI converge around five key values: fairness, transparency, autonomy, nonmaleficence and accountability. The principles of beneficence and sustainability were advanced, although less frequently highlighted. Despite a reassuring level of convergence across policy documents, the practical implementation of these ethical principles in health care and particularly for skin-specific AI tools remains unclear. Furthermore, evidence of real-world integration of these ethical frameworks into developing AI-based skin cancer diagnostic tools is lacking. Future work seeks to address this gap by analysing commercially available skin cancer diagnostic applications using predefined ethical themes to identify case-specific ethical issues and to evaluate the extent to which these issues are addressed within the existing ethical frameworks. This will provide guidance for developers and adopters of these tools, ensuring responsible design and deployment.

Identifying presence of cybersickness symptoms using AI-based predictive learning algorithms

Cybersickness (CS) affects a large proportion of virtual reality (VR) users causing a combination of nausea, headaches and dizziness which would create barriers to the users, VR designers/developers and the stakeholders in the production industry. Although design principles suggest methods to avoid CS, challenges remain as new demands and systems continue to penetrate the competitive market. The dilemma is whether to use VR technology by experiencing the ultimate virtual world using a head-mounted display (HMD) with possible CS triggers or to avoid the triggers by avoiding using VR. With the huge success and potential in the entertainment industry, it is very important to focus on the solutions to handling CS dilemmas. Therefore, the main observation for the developers is to have a guide around the set of established design principles aiming to broadly reduce CS. In this paper, we provide a method to apply artificial intelligence (AI) techniques and use machine learning (ML) algorithms including support vector machines (SVMs), decision trees (DTs) and K-nearest neighbours (KNNs) to predict CS outcomes. Based on our findings, we have observed that DT and SVM surpassed KNN in test accuracy. Additionally, DT exhibited better results than both SVM and KNN in train accuracy. By exploiting the power of ML, developers will be able to predict the potential occurrence of CS while developing VR projects to find ways to alleviate CS more effectively.

Stroke Risk Prediction Using Artificial Intelligence Techniques Through Electronic Health Records

Nowadays, Electronic Health Records (EHR) include critical information in the text format. In order to make medical decisions more efficient, the text should be processed and code deliberated. In this report, we applied Artificial Intelligence (AI) techniques to improve stroke risk prediction based on the EHR text. The system based on Natural Language Processing (NLP) generates structured text from EHR, followed by applying Machine Learning (ML) techniques to classify the text as a "good" or "bad" indicator, which is used for prediction. The ML models here we used include logistic regression and Support Vector Machine (SVM). Our results show that both models can classify the text precisely and make predictions accurately.

System Evaluation of Artificial Intelligence and Virtual Reality Technology in the Interactive Design of Interior Decoration

Applying artificial intelligence (AI) and virtual reality (VR) technology to interior decoration design can effectively shorten the time of communication between customers and the designers, the design time, and the distance between designers, customers, and the space to be designed, which meet contemporary needs. This paper aimed to study how to analyze and design an interactive interior decoration design system based on AI and VR technology. This paper also used the fuzzy comprehensive evaluation (FCE) method to comprehensively verify and evaluate the designed system. According to the system verification experiment in this paper, the sensitivities of the left, right, front, and rear operating handles for 50 times were 99.38%, 99.36%, 99.49%, and 99.21%, respectively. In addition, when the number of users simultaneously using the system was 60, the system’s stability, security, response time, and stuck time were 98.72%, 98.40%, 0.834 s, and 0.322 s, respectively. Based on a series of evaluative tests, the interactive design system for interior decoration presented in this paper is feasible and worthy of further promotion and application. The system designed in this article has reference value for the academic community and boasts certain innovative aspects.

Cost supervision mining from EMR based on artificial intelligence technology.

To effectively monitor medical insurance funds in the era of big data, the study tries to construct an inpatient cost rationality judgement model by designing a virtuous cycle of inpatient cost supervision information system and exploring a complete set of inpatient cost supervision methods. To lay the foundation for applying artificial intelligence (AI) technology in medical insurance cost control supervision and provide feasible paths and available tools for medical insurance cost control managers. By way of collecting and cleaning electronic medical record (EMR) data from 2016 to 2018 of a city in East China, focusing on basic patient information and cost information, and using a combination of machine learning modeling and information system construction, the study tries to form a feasible inpatient cost supervision method and operation path. The set of the regulatory method, applied in nursing homes of a city in East China, is compelling. The accuracy rates of rationality judgement in different main diseases are stable up to 80%, the false positive rate is steady within 10%, and rehabilitation fee days of hospitalization, and the number of complications are important factors affecting the rationality of the inpatient cost. The model construction and optimization method combining machine learning and information system can make practical cost rationality judgement on medical institution's inpatient cost data, which can directly reflect the key influencing factors of relevant inpatient costs, and achieve the effect of guiding medical behavior and improving the efficiency of medical insurance fund use.

Artificial intelligence research in the COVend COVID-19 clinical trial project

The COVend project aims at delivering a new effective therapy, FX06, against the SARS-CoV-2 virus infection for the management of the COVID-19 disease in hospitals. Nine of the 17 partners of the project consortium are hospitals responsible for collecting study subjects and administering the FX06 therapy to the patients. Although the clinical trial (IXION) has the main role in the project, the project has also a work package which develops and applies artificial intelligence (AI) methods to the data collected during the 28-day study period from the patients receiving the therapy. The AI work package applies exploratory data analysis methods to find patterns and profiles of the patients. Combined with the data about treatment methods and patient outcomes, the aim is to provide decision support for the therapy intervention in the later stage of the project.

Retinal age as a predictive biomarker of the diabetic retinopathy grade

ObjectiveTo apply artificial intelligence (AI) techniques, through deep learning algorithms, for the development and optimization of a system for predicting the age of a person based on a color retinography and to study a possible relationship between the evolution of retinopathy diabetes and premature ageing of the retina. MethodsA convolutional network was trained to calculate the age of a person based on a retinography. Said training was carried out on a set of retinographies of patients with diabetes previously divided into three subsets (training, validation and test). The difference between the chronological age of the patient and the biological age of the retina was defined as the retinal age gap. ResultsA set of 98,400 images was used for the training phase, 1000 images for the validation phase and 13,544 for the test phase.The retinal gap of the patients without DR was 0.609 years and that of the patients with DR was 1905 years (p < 0.001), with the distribution by degree of DR being: mild DR: 1541 years, moderate DR: 3017 years, DR severe: 3117 years and proliferative DR: 8583 years. ConclusionsThe retinal age gap shows a positive mean difference between diabetics with DR versus those without DR, and it increases progressively, according to the degree of DR. These results could indicate the existence of a relationship between the evolution of the disease and premature ageing of the retina.

Artificial Intelligence, Creativity, and Intentionality: The Need for a Paradigm Shift

Artificial Intelligence, Creativity, and Intentionality: The Need for a Paradigm Shift

Abstract 5381: A broad-use deep learning model based on multi-dimensional morphology to identify and characterize tumor cell heterogeneity

Abstract Technological advances have ushered in a new multi-omics era that has elucidated more granular information on cell types and functions at single-cell resolution. Integrating morphology as a multi-dimensional readout of cell identity, state, and function is an area of broad interest, since it has been historically difficult to quantify morphological properties. Here, we apply artificial intelligence (AI) and multi-dimensional morphology to characterize a variety of human cell types and states using the Deepcell platform. Captured brightfield images of cells flowing inside a microfluidic channel are analyzed by AI models in real-time and cells of interest can be sorted for downstream molecular or functional analyses. We developed an AI model, termed ‘Human Foundation Model’ (HFM), for broad use that combines machine learning with computer vision for multi-dimensional morphology analysis to provide interpretable features on a broad range of human cells and diverse sample types. The HFM training and validation datasets, comprising 1.18 million and 4.05 million cell images, respectively, were chosen to capture a comprehensive range of visual features and cover a wide spectrum of biological diversity. Samples included 6μm polystyrene beads (as controls) and multiple immune and carcinoma cell lines. Subsets of these images were labeled for specific cell features then included in the supervised portion of the model training. We apply the HFM as part of the Deepcell system to identify and characterize the morphological heterogeneity of tumor cells in different types of cancers such as melanoma, non-small cell lung carcinoma, and malignant pleural effusions. We further demonstrate the ability of the platform to distinguish different cell types (epithelial, stromal, immune, and endothelial cells) commonly found in tumors that typically require complex antibody panels. With the Deepcell platform and the HFM, we determined cell identities using morphological features associated with distinct processes characterized by granules, vesicles, cell size, pigmentation and others. Furthermore, sorted cells of interest are unlabeled and can be retrieved for further analysis. This AI-powered technology can be applied to diverse areas including drug screenings/mechanism of action, tumor biology, and functional genomics. Citation Format: Stephane C. Boutet, Anastasia Mavropoulos, Andreja Jovic, Jeanette Mei, Nianzhen Li, Kiran Saini, Senzeyu Zhang, Chassidy Johnson, Vivian Lu, Ryan Carelli, Kevin B. Jacobs, Mahyar Salek, Maddison Masaeli. A broad-use deep learning model based on multi-dimensional morphology to identify and characterize tumor cell heterogeneity. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5381.

A study on the optimal condition of ground truth area for liver tumor detection in ultrasound images using deep learning

PurposeIn recent years, efforts to apply artificial intelligence (AI) to the medical field have been growing. In general, a vast amount of high-quality training data is necessary to make great AI. For tumor detection AI, annotation quality is important. In diagnosis and detection of tumors using ultrasound images, humans use not only the tumor area but also the surrounding information, such as the back echo of the tumor. Therefore, we investigated changes in detection accuracy when changing the size of the region of interest (ROI, ground truth area) relative to liver tumors in the training data for the detection AI.MethodsWe defined D/L as the ratio of the maximum diameter (D) of the liver tumor to the ROI size (L). We created training data by changing the D/L value, and performed learning and testing with YOLOv3.ResultsOur results showed that the detection accuracy was highest when the training data were created with a D/L ratio between 0.8 and 1.0. In other words, it was found that the detection accuracy was improved by setting the ground true bounding box for detection AI training to be in contact with the tumor or slightly larger. We also found that when the D/L ratio was distributed in the training data, the wider the distribution, the lower the detection accuracy.ConclusionsTherefore, we recommend that the detector be trained with the D/L value close to a certain value between 0.8 and 1.0 for liver tumor detection from ultrasound images.

Wavelet transforms for feature engineering in EEG data processing: An application on Schizophrenia

Applying Artificial Intelligence (AI) in the healthcare domain is getting benefitted day by day with the advancement of approaches, one of them being Bio-Signal analysis. In Bio-signals, efficient feature engineering and feature extraction (FE) is necessary for optimal results. Features can be extracted from different methods by Time, Frequency, and Time-frequency domains. Time-frequency domain features are the most advanced and perform well for most AI-based signal analysis problems. We introduced the application of Wavelet Scattering Transform (WST) to neuro-disorder classification and provided a comparative study with Continuous Wavelet Transform (CWT) and Discrete Wavelet Transform (DWT) for schizophrenia disease classification. We are one of the first to apply WST to EEG data for classifying neurological disorders. We have also extracted 12 statistical features from the data before sending them to classifiers for classification. We built six Machine Learning (ML) algorithms from two categories core/traditional ML (Logistic regression and Support vector machine) and Ensemble Learning (EL) (Decision Trees, Random Forest, AdaBoost, and Gradient Boost). In total we have conducted 18 experiments, our study found that ensembling methods performed better when features are extracted from CWT and DWT. At the same time, traditional ML methods performed better than EL methods when features are extracted from WST. Overall SVM performed better, but the best results are attained by Decision trees which are; 97.98%; 98.2%;97.72%; 95.94; values of accuracy, sensitivity, specificity, and Kappa score respectively, and execution time of 48.04 s; our proposed method performed better than the reported state-of-the-art methods.

Predicting Success for Web Product through Key Performance Indicators based on Balanced Scorecard with the Use of Machine Learning

Machine Learning (ML) can be proved as an important tool in planning better business strategies. For the purposes of the present study, the prospect for the development of an electronic platform by a technology firm providing financial services is explored. The purpose of this article is to demonstrate the ways in which a start-up can predict the success of an online platform prior to its market launch. The prediction is achieved by applying Artificial Intelligence (AI) on Key Performance Indicators (KPIs) derived from the customers’ perspective, as shown in the Balanced Scorecard (BSC). The research methodology was quantitative and online questionnaires were used to collect empirical quantitative data related to bank loans. Subsequently, KPIs were created based on the collected data, to measure and assess the success of the platform. The effectiveness of the model was calculated up to 91.89%, and thus, it is estimated that the online platform will be of great success with 91.89% validity. In conclusion, prediction was found to be crucial for businesses to prevent a dire economic situation. Finally, the necessity for businesses to keep up with technological advances is highlighted.

Human Activity Detection Using Smart Wearable Sensing Devices with Feed Forward Neural Networks and PSO

Hospitals must continually monitor their patients’ actions to lower the chance of accidents, such as patient falls and slides. Human behavior is difficult to track due to the complexity of human activities and the unpredictable nature of their conduct. As a result, creating a static link that is used to influence human behavior is challenging, since it is hard to forecast how individuals will think or act in response to a certain event. Mobility tracking depends on intelligent monitoring systems that apply artificial intelligence (AI) applications referred to as “categories”. Because motion sensors, such as gyroscopes and accelerometers, output unconnected data that lack labels, event detection is a vital task. The fall feature parameters of tridimensional accelerometers and gyroscope sensors are presented and used, and the classification technique is based on distinguishing characteristics. This study focuses on the age-old problem of tracking turbulence in motion to improve detection precision. We trained the model, considering that detection accuracy is limited by factors such as the subject’s mass, velocity, and gait style. This is performed by employing an experimental dataset. When we used the sophisticated technique of particle swarm optimization (PSO) in combination with a four-stage forward neural network (4SFNN) to forecast four different types of turbulent motion, we observed that the total prediction accuracy was 98.615% accurate.

Predicting Gross Motor Function in Children and Adolescents with Cerebral Palsy Applying Artificial Intelligence Using Data on Assistive Devices

Data obtained from routine clinical care find increasing use in a scientific context. Many routine databases, e.g., from health insurance providers, include records of medical devices and therapies, but not on motor function, such as the frequently used Gross Motor Function Measure-66 (GMFM-66) score for children and adolescents with cerebral palsy (CP). However, motor function is the most common outcome of therapeutic efforts. In order to increase the usability of available records, the aim of this study was to predict the GMFM-66 score from the medical devices used by a patient with CP. For this purpose, we developed the Medical Device Score Calculator (MDSC) based on the analysis of a population of 1581 children and adolescents with CP. Several machine learning algorithms were compared for predicting the GMFM-66 score. The random forest algorithm proved to be the most accurate with a concordance correlation coefficient (Lin) of 0.75 (0.71; 0.78) with a mean absolute error of 7.74 (7.15; 8.33) and a root mean square error of 10.1 (9.51; 10.8). Our findings suggest that the MDSC is appropriate for estimating the GMFM-66 score in sufficiently large patient groups for scientific purposes, such as comparison or efficacy of different therapies. The MDSC is not suitable for the individual assessment of a child or adolescent with CP.

Efficient secure data storage based on novel blockchain model over IoT-based smart computing systems

In recent years, edge-related smart computing is the way to process and store data via outsourcing environments. Security is the main progressive concept in smart-related Internet of Things (IoT) software denied networks to share and increase the scalability and efficiency in data storage. Because of the rapid growth of different smart services, different security-related problems may appear in resource processing and sharing data in real-time computing systems. To increase the reliability in secure data storage and satisfy the basic requirements related to IoT-related smart computing networks. So propose and implement a Novel Artificial Intelligence based Blockchain Secure Model (NAIBSM) to provide efficient secure data storage in IoT-related smart computing systems. This model flexibly captures each user authentication for the detection of different user-related attacks (i.e. distributed denial-of-service (DDOS)) in the storage of data via applying Artificial Intelligence (AI) calculation method i.e. Leverage Bat algorithm to explore complex features. Build a blockchain at the server side to provide secure communication and reliability of storing data on the terminal of IoT. This approach provides random hash values to each data to ensure blockchain for the integrity of data and uses a weight-based data storage procedure to arrange/store and classify data to each user with secure and unsecured storage in smart computing networks. The experimental results of the proposed model are to explore complex security features and ensure the performance of secure authentication to each user and better security, accuracy, and low communication overhead in IoT-related smart computing data storage and sharing systems.

Machine Learning Identification of Obstructive Sleep Apnea Severity through the Patient Clinical Features: A Retrospective Study.

To evaluate the role of clinical scores assessing the risk of disease severity in patients with clinical suspicion of obstructive sleep apnea syndrome (OSA). The hypothesis was tested by applying artificial intelligence (AI) to demonstrate its effectiveness in distinguishing between mild-moderate OSA and severe OSA risk. A support vector machine model (SVM) was developed from the samples included in the analysis (N = 498), and they were split into 75% for training (N = 373) with the remaining for testing (N = 125). Two diagnostic thresholds were selected for OSA severity: mild to moderate (apnea-hypopnea index (AHI) ≥ 5 events/h and AHI < 30 events/h) and severe (AHI ≥ 30 events/h). The algorithms were trained and tested to predict OSA patient severity. The sensitivity and specificity for the SVM model were 0.93 and 0.80 with an accuracy of 0.86; instead, the logistic regression full mode reported a value of 0.74 and 0.63, respectively, with an accuracy of 0.68. After backward stepwise elimination for features selection, the reduced logistic regression model demonstrated a sensitivity and specificity of 0.79 and 0.56, respectively, and an accuracy of 0.67. Artificial intelligence could be applied to patients with symptoms related to OSA to identify individuals with a severe OSA risk with clinical-based algorithms in the OSA framework.

Applied Artificial Intelligence in Manufacturing and Industrial Production Systems: PEST Considerations for Engineering Managers

Presently, artificial intelligence (AI) is playing a leading role in our contemporary world via numerous applications. Despite its many advantages, analytical frameworks highlighting the implications of AI applications are still evolving. Particularly, in manufacturing and industrial production where novel technologies are continuously being harnessed. Consequently, AI and the implications of its applications have relatively remained a grey area for many engineering managers who are key players in the gravitation of manufacturing and industrial production towards the fourth industrial revolution and more recently, the fifth industrial revolution, generally termed as Industry 4.0 (I4.0) and Industry 5.0 (I5.0), respectively. In this study, the implications of AI applications in the general context of manufacturing and industrial production, are presented to provide insight for engineering managers. These implications are discussed via PEST ( <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</u> olitical, <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e</u> conomic, <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</u> ocial, and <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">t</u> echnological) considerations of the broad implications of the adoption of AI techniques in manufacturing and industrial production systems. A new engineering management model has not been proposed in this paper. Rather, a discussion aimed at serving as a tool for the appraisal of the implications of the general applications of AI by engineering managers, who may not be AI specialists or data science experts is presented.

Abstract P5-04-03: Analyzing the performance of Thermalytix, an AI-based breast cancer screening solution, in a community setting

Abstract Background: Despite improvements in treatment strategies, breast cancer survival rates remain low in India due to a lack of awareness and late stage of presentation. If diagnosed and treated early, breast cancer survival rates improve. Screening mammography, the gold standard in cancer diagnosis, is not feasible in a resource-constrained setting. Niramai’s novel breast cancer screening technology, Thermalytix™, applies Artificial Intelligence (AI) over thermography, to give an automated interpretation of the breast thermal images. The portable Thermalytix test has been so far used to screen 60000 women in community settings across India and Kenya. This study is a recent evaluation of the test in rural setting. Methods: Women who provided written consent and who underwent Thermalytix tests in community-based screening camps at primary health centers (PHCs) in Afzalpur Taluk of Gulbarga District, Karnataka, India between 01 August 2021 to 15 June 2022 were included in this study. Five thermal images in multiple views were analyzed using Niramai’s patented algorithm. Automated analysis of the thermal images produced a screening report and triaged the participants for follow-up. In case of abnormal thermal activity, Thermalytix triaged women as ‘red’ and were referred to the district hospital for follow-up with breast ultrasound and/or other investigations and and were recorded into the following three categories: Normal (BI-RADS 1), Abnormal - benign (BI-RADS 2/3) and Abnormal - malignant (BI-RADS 4/5). If no abnormal thermal patterns were detected by Thermalytix, women would be recommended routine screening. Findings: The analysis included 3531 women were included in the analysis and the median age in the cohort was 42 years. Of them, 97 (2.74%) women were triaged ‘red’ by Thermalytix indicating a suspicion of breast abnormality. As on 15 June 2022, 29 (30%) out of 97 women underwent standard follow-up investigations of which two cases of carcinoma breast, one case of phyllodes, one case of tuberculosis mastitis and seven other benign cases were identified, indicating that Thermalytix has a positive predictive value of 35.71% (11/29) in detecting benign and malignant breast lesions. Furthermore, a Patient experience questionnaire was used to asses their experience. 98.71% women were being screened for breast cancer for the first time in their lives. 93.9% women said they were very satisfied with Thermalytix screening experience and remaining 6.1% said they were satisfied, thus aiding in strong acceptability and adoption of the test. Interpretation: In resource-constrained settings such as India, where less than 2% of the women in the country have ever got screening for breast cancer, the portable, no-radiation Thermalytix test is an accessible breast cancer screening solution. Thermalytix’s patient-friendly features - privacy-aware, painless, comfortable, and radiation-free make it favorable for population screening in India, and thus, can increase the uptake of breast cancer screening. With the device’s capacity to fit into a backpack, it can make breast cancer screening accessible to even remote areas with limited resources. However, following up on women who were found to be Thermalytix positive still remains a challenge. Future studies will emsure that follow-up after abnormal breast screening is part of the approved clinical protocol. Results of the screening program 3 malignancies and 8 benign lesions found in 29 Thermalytix RED patients Citation Format: Geetha Manjunath, Lakshmi Krishnan, Gargi Deshpande, Purnima Madhivanan, Karl Francis Krupp. Analyzing the performance of Thermalytix, an AI-based breast cancer screening solution, in a community setting [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P5-04-03.

I Asked a ChatGPT to Write an Editorial About How We Can Incorporate Chatbots Into Neurosurgical Research and Patient Care….

I Asked a ChatGPT to Write an Editorial About How We Can Incorporate Chatbots Into Neurosurgical Research and Patient Care….

Sugarcane Disease Detection Using CNN-Deep Learning Method: An Indian Perspective

Agriculture produce especially sugarcane crop is no exception to diseases as compared to the other crops. Sugarcane diseases can be mitigated more successfully if they are identified properly and in the early stages of the diseases. Disease prone sugarcane crop directly affects the production quality and quantity. Sugarcane infections are a cause of worry for the farmers because they can wipe out the entire crop field, causing financial loss. Researchers are working on applying Artificial Intelligence (AI) techniques, like Machine Learning (ML) and Deep Learning (DL), to analyse the agricultural data (yield prediction, selling price forecasting, climate, and soil quality etc.) and prevent crop damage due to various reasons, diseases being one of them. Sugarcane producing farmers are required to be enriched with real-time data analysis using various computational techniques along with managing huge datasets. Deep neural network which includes Convolutional Neural Network (CNN) is a modern technique for agricultural disease detection. Hence, this paper presents the feasibility study and the effectiveness of DL based CNN algorithm in the disease detection of crops with special reference to selective four diseases of sugarcane crop in India. The study was prompted by the rapid evolution of sugarcane disease classes and farmers' lack of disease diagnostic and recognition skills. To solve this challenge, deep learning and computer vision are adopted. By categorizing sugarcane images into two groups: healthy and unhealthy/diseased (with disease type), the trained model is able to fulfill its goal. Using a simple CNN with four discrete classes, the analysis shows an accuracy of 98.69% for sugarcane disease detection. Further to guide the farmers, a web-based application is developed for sugarcane crop disease detection and implemented to fetch and monitor the data about the diseases. The paper also puts forth a few future research areas such as: (i) the user can enter the feedback which can be tuned with the model for accurately predicting the sugarcane crop diseases and dynamic database update; (ii) the effect of disease detection can be combined with agricultural productivity enhancement and price forecasting using AI tools and techniques, thereby aiding the farmers in effective decision making.