Title A Mini Project Report submitted to Manipal Institute of Technology in partial fulfillment of the requirements for the award of the degree of BACHELOR OF TECHNOLOGY IN MECHATRONICS Submitted by Dhee Choudhary - 200929270 Ravi Surya - 200929266 Sidharth Nohwar - 200929108 Anurag Jaiswal - 200929076 Rohit Milan - 200929010 Under the guidance of Dr. Umesh Kumar Sahu Assistant Professor DEPARTMENT OF MECHATRONICS MANIPAL INSTITUTE OF TECHNOLOGY (A Constituent of Manipal Academy of Higher Education) MANIPAL - 576104, KARNATAKA, INDIA November 2023 ABSTRACT Fabric, a testament to human ingenuity, has evolved from handmade textiles to modern electronic fabrics. Within the intricate realm of textile manufacturing, the weave pattern stands as a fundamental element, significantly impacting aesthetics, structural integrity, and overall fabric appeal. Recognizing and analyzing these intricate weave patterns is vital for efficient textile production. Unfortunately, current methods primarily rely on labor-intensive manual inspection, introducing inefficiencies and subjectivity, ultimately affecting recognition accuracy. This project aims to develop an automated fabric pattern recognition system utilizing the Radon Transform and a machine learning model like AlexNet. By automating this critical task, the project seeks to enhance the efficiency and precision of textile production while alleviating the subjective human factors that often impede consistent results. The system promises to revolutionize textile manufacturing by reducing errors and stress for human inspectors and meeting the diverse demands of discerning customers. Textile manufacturers and quality control teams stand to benefit significantly, as the system offers the potential to usher in a new era of high-quality textile production. 2 LIST OF FIGURES 1.1 Project Work Timeline 2 3 LIST OF TABLES 1.1 Individual Objectives of project work 4 2 TABLE OF CONTENTS Abstract 2 List of Figures 3 List of Tables 4 Chapter 1 1 Introduction 1.1 Introduction 1 1.2 Motivation 1 1.3 Project Work Schedule 1 1.4 Individual Objectives of project work 1 Chapter 2 Literature Review and Theoretical Background 3 2.1 Review of Literature 3 2.2 Summary of Literature Review 3 2.3 Theoretical Background 3 Chapter 3 Problem Definition and Objectives 4 3.1 Problem Definition 4 3.2 Objectives 4 Chapter 4 5 Methodology 4.1 Experiment and Implementation 5 4.2 Summary 5 Chapter 5 6 Results and Discussion 5.1 Results and Discussion 6 5.2 Discussion on Results 6 Chapter 6 Conclusion and Future Scope 7 6.1 Conclusion 7 6.2 Scope for Future Work 7 5 References 9 Annexures 11 Plagiarism Check 12 6 CHAPTER 1 Introduction 1.1 Introduction The realm of textile manufacturing has been a cornerstone of human innovation and craftsmanship for centuries. The creation of fabric is an intricate process, encompassing a rich tapestry of history, culture, and technological progress. One fundamental element that distinguishes fabrics and holds immense significance is the weave pattern. Weave patterns not only define the visual aesthetics of a textile but also play a pivotal role in determining its structural integrity and overall appeal. The art of recognizing and analyzing intricate weave patterns is a crucial step that precedes further processing of textiles through weaving machinery. In the present era, this process predominantly relies on labor-intensive manual inspection performed by human experts. These experts often use tools like microscopes and magnifying glasses to scrutinize fabric patterns. However, such manual processes come with inherent inefficiencies, subjectivity, and the risk of human factors such as stress and fatigue, all of which can considerably impact the accuracy and consistency of recognition results. This traditional approach, while having served well over the years, faces numerous challenges in today's fast-paced and quality-centric industries. 1.2 Labor-Intensive Processes: Manual inspection processes are time-consuming and require skilled human labor, resulting in high operational costs and potential bottlenecks in production. Subjectivity: Human inspectors introduce subjective judgment, leading to variability in recognition results. Factors such as lighting conditions and personal biases can affect the accuracy and consistency of pattern recognition. Inefficiency: The traditional methods are limited in their capacity to handle the diverse and complex patterns found in modern textiles. As the textile industry evolves, so too must its inspection methods to meet the demands of the market. Motivation The motivation behind this project is rooted in the need for innovation and automation in the textile manufacturing industry. It aims to address the aforementioned challenges and inefficiencies associated with traditional fabric pattern recognition methods. By automating this critical task, we seek to enhance the efficiency and precision of textile production, ultimately resulting in the production of consistent, high-quality fabrics. This project is motivated by the following factors: Enhancing Efficiency: The automation of fabric pattern recognition will significantly reduce the time and labor required for inspections, leading to cost savings and improved production efficiency. 1 Improving Accuracy: Automation eliminates human subjectivity and ensures a higher degree of accuracy in pattern recognition, reducing errors in the manufacturing process. Meeting Customer Demands: The textile industry is driven by consumer demand for a wide range of intricate and customized fabric patterns. Automation enables quicker adaptation to market trends and customer preferences. Reducing Stress and Fatigue: By relieving human inspectors of repetitive and strenuous tasks, the project aims to improve the work environment and job satisfaction. 1.3 Project Work Schedule 1.3.1 Data Preparation and Preprocessing Week 1: Dataset Acquisition Objective 1: Identify and collect the dataset of cloth images. Objective 2: Implement dataset searching to ensure dataset diversity. Week 1: Data Preprocessing Objective 3: Develop and apply a preprocessing pipeline for image standardization. Objective 4: Address challenges related to lighting and unclear textures. 1.3.2 Radon Signature Extraction Week 2: Radon Transform Objective 5: Apply the Radon Transform to preprocessed images. Objective 6: Normalize Radon Signatures using the L2-norm. 1.3.3 Machine Learning Model Development Week 3: Model Selection Objective 7: Research and select an appropriate machine learning model (e.g., CNN). Objective 8: Design the architecture and prepare the model for training. Week 3: Model Training and Fine-Tuning Objective 7: Train the selected model using the Radon Signature dataset. Objective 8: Fine-tune the model's hyperparameters to optimize performance. 1.3.4 Classification and Evaluation Week 3: Classification Framework Objective 9: Develop a classification framework to apply the trained model to test images. Objective 10: Implement evaluation metrics for accuracy, precision, and recall. 2 Week 3: Testing and Evaluation Objective 9: Test the model's performance on unseen cloth images. Objective 10: Evaluate the model's classification accuracy and precision. 1.3.5 Documentation and Reporting Week 4: Report Writing 1.4 Summarize project objectives, methodologies, and results in a report. Objectives of project work The project's objectives are delineated into specific tasks, each contributing to the overall goal of automated fabric pattern recognition. These objectives encompass the entire process from image acquisition to pattern classification, with a focus on Radon Signature-based recognition and machine learning: 1.4.1 Image Acquisition and Dataset Searching Objective 1: Identify and acquire a diverse and representative dataset of cloth images. This dataset should encompass a wide range of fabric patterns, textures, and variations. Objective 2: Implement a systematic dataset searching process to ensure that the collected images cover various fabric types and patterns used in real-world scenarios. The dataset should be well-structured, labeled, and suitable for training and testing. 1.4.2 Preprocessing Objective 3: Develop a preprocessing pipeline to standardize the acquired images. This includes tasks such as resizing the images, converting them to grayscale, and applying edge detection algorithms, ensuring that images are prepared for Radon Transform. Objective 4: Address any challenges related to inconsistent lighting conditions and unclear textures in the images through appropriate preprocessing techniques. 1.4.3 Radon Signature Extraction Objective 5: Apply the Radon Transform to the preprocessed images at various angles, extracting the Radon Signature for each image. This step involves computing the variance of the Radon Transform values along different angles. Objective 6: Normalize the Radon Signatures using the L2-norm to make them invariant to the scale of the images. 1.4.4 Machine Learning Objective 7: Train a machine learning model, such as a Convolutional Neural Network (CNN) like AlexNet, using the dataset of Radon Signatures. The model should be designed to recognize fabric patterns based on the Radon Signatures. Objective 8: Fine-tune the machine learning model as needed to optimize its performance and generalization to various fabric patterns. 3 1.4.5 Classification Objective 9: Develop a classification framework to apply the trained machine learning model to unseen cloth images. The model should predict the fabric pattern class based on the Radon Signatures of the test images. Objective 10: Evaluate the classification accuracy, precision, and recall of the model to assess its performance in real-world applications. This structured approach divides the project into distinct phases, each with specific objectives and tasks, to achieve the overarching goal of automated fabric pattern recognition 4 CHAPTER 2 Literature Review and Theoretical Background 2.1 Review of Literature The textile industry has witnessed a significant transformation in recent years, with automation and robotics playing a pivotal role in enhancing efficiency and precision in various processes, including fabric pattern recognition. The automation of fabric pattern recognition involves a combination of computer vision, image processing, and machine learning techniques. Recent research has seen advancements in these areas to improve the quality and accuracy of pattern analysis. Key developments include: 1. Preprocessing Techniques: Recent research has focused on developing robust preprocessing methods to standardize images, addressing challenges related to lighting conditions and unclear textures. These techniques ensure that images are well-suited for subsequent pattern analysis. 2. Texture Analysis: Texture-based methods have been a staple in fabric pattern recognition. Recent research has explored advanced texture analysis approaches to handle complex fabric patterns, including the use of deep learning models for feature extraction. 3. Edge Detection and Segmentation: Improved edge detection and segmentation algorithms have emerged to enhance the extraction of fabric patterns from images. These techniques help in delineating pattern boundaries accurately. Machine learning, particularly deep learning, has become a cornerstone of fabric pattern recognition. Recent developments in this area include: 1. Convolutional Neural Networks (CNNs): CNNs have gained prominence in the textile industry for their ability to recognize intricate patterns. Recent studies have explored variations of CNN architectures to optimize pattern recognition accuracy. 2. Transfer Learning: Researchers have leveraged transfer learning techniques, using pretrained models on large image datasets to improve the performance of fabric pattern recognition models, especially in cases with limited labeled data. 3. Fine-Tuning and Hyperparameter Optimization: Recent research emphasizes the importance of fine-tuning and optimizing hyperparameters to tailor machine learning models to specific fabric pattern recognition tasks. 2.2 Summary of Literature Review While substantial progress has been made in automating fabric pattern recognition, the existing literature underscores the need for addressing the unique challenges posed by diverse fabric types and varying lighting conditions. These challenges remain areas where further research and innovation are crucial to achieving a comprehensive and robust automated fabric pattern recognition system. Specifically, the current literature indicates the following research gaps: 5 CHAPTER 2 1. Fabric Type Adaptability: Most existing methods are tailored to specific fabric types, and there's a gap in research that explores techniques capable of adapting to a broad range of textiles. Developing methods that can accommodate various fabric types would enhance the versatility of automated recognition systems. 2. Lighting Variability: The impact of varying lighting conditions on pattern recognition accuracy remains a challenge. Research focusing on strategies to make recognition systems more robust to lighting effects is necessary to ensure consistent performance across different environments. 3. Human-Machine Collaboration: An emerging trend emphasizes the synergy between human expertise and automated recognition systems. Yet, more research is needed to define the optimal balance and interaction between human inspectors and machines, creating efficient, reliable, and user-friendly systems. Addressing these research gaps will contribute to the further refinement and expansion of automated fabric pattern recognition within the textile industry, enabling the development of more comprehensive and adaptable systems. 2.3 Theoretical Background The theoretical background for automated fabric pattern recognition centers on the design task, which comprises critical components: image preprocessing, feature extraction, pattern classification, and integration with human expertise. Image preprocessing encompasses resizing, grayscale conversion, edge detection, and lighting correction to prepare images for analysis. Feature extraction involves the use of the Radon Transform, texture analysis, and Convolutional Neural Networks (CNNs) to capture informative features from fabric images. Pattern classification involves selecting a machine learning model, such as a CNN, and finetuning it to classify fabric patterns. Integration with human expertise focuses on creating a user-friendly interface and ensuring the system collaborates with human inspectors. Key programming parameters, including Radon Transform settings and CNN hyperparameters, are meticulously chosen. The Radon Transform, with its angle-based approach, and CNNs, renowned for their hierarchical feature extraction, are selected for their suitability in fabric pattern recognition. The integration of automated systems with human expertise enhances flexibility and reliability. This comprehensive framework provides the foundation for the development of an effective fabric pattern recognition system, ensuring technical depth and breadth in methodology selection 6 and design. CHAPTER 3 Problem Definition and Objectives 3.1 Problem Definition The textile manufacturing industry, known for its intricate and diverse fabric patterns, has traditionally relied on labor-intensive manual inspection for fabric pattern recognition. This manual process, performed by human experts equipped with tools such as microscopes and magnifying glasses, presents several challenges and limitations. The problems necessitating the initiation of the present mini-project work are as follows: 1. Labor-Intensive Inspection: Traditional fabric pattern recognition heavily depends on the expertise of human inspectors. Manual inspection is time-consuming, laborintensive, and prone to human errors, leading to inefficiencies in the manufacturing process. 2. Subjective Human Factors: Human inspectors, even with advanced tools, can be affected by subjective factors such as stress, fatigue, and varying levels of expertise. This subjectivity can introduce variations and inconsistencies in the recognition results. 3. Limited Scalability: As the textile industry experiences growing demands for unique and intricate fabric patterns, manual inspection becomes increasingly challenging to scale. This limitation hampers the industry's ability to meet diverse customer preferences efficiently. 4. Quality Control Challenges: Manual inspection may not always detect subtle defects or discrepancies in fabric patterns, leading to inconsistencies in the quality of manufactured textiles. 5. Reduced Cost Efficiency: Labor-intensive manual inspection incurs substantial labor costs for the industry. Finding cost-efficient ways to maintain quality while reducing these expenses is a pressing concern. In light of these challenges, the present mini-project work aims to develop an automated fabric pattern recognition system that mitigates the limitations of manual inspection. By leveraging computer vision, image processing, and machine learning techniques, this project seeks to provide a scalable, objective, and cost-efficient solution for the recognition of intricate fabric patterns, addressing the diverse demands of the textile industry and enhancing the quality control process. The objectives of this project work are designed to tackle these problems and contribute to the automation and optimization of fabric pattern recognition. 3.2 Objectives 1. Preprocessing Pipeline Design Type: Design Objective: Develop a comprehensive preprocessing pipeline to standardize fabric images, incorporating resizing, grayscale conversion, edge detection, and lighting correction. Ensure that the preprocessing steps effectively prepare images for pattern recognition. 7 CHAPTER 3 2. Feature Extraction and Analysis Type: Analysis Objective: Investigate and implement advanced feature extraction methods, such as the Radon Transform, texture analysis, and Convolutional Neural Networks (CNNs). Analyze the performance of these techniques in capturing distinctive fabric patterns from images. 3. Machine Learning Model Selection and Optimization Type: Design and Experimental Work Objective: Select a suitable machine learning model for fabric pattern recognition, with a particular emphasis on CNNs. Optimize the model's architecture and fine-tune hyperparameters to enhance its accuracy and efficiency in pattern classification. 4. Integration with Human Expertise Type: Analysis and Design Objective: Investigate methods for effectively integrating the automated system with human expertise in fabric pattern recognition. Analyze how the system can complement and enhance human inspectors' capabilities and decision-making. 5. Programming and Parameter Optimization Type: Experimental Work Objective: Implement the designed preprocessing pipeline, feature extraction methods, and machine learning model in code. Optimize programming parameters, including Radon Transform settings, CNN hyperparameters, and user interface functionalities. 6. Performance Validation and Evaluation Type: Validation and Experimental Work Objective: Conduct rigorous testing and validation of the automated fabric pattern recognition system. Evaluate its performance in terms of recognition accuracy, precision, recall, and computational efficiency. Address any identified issues and limitations. 7. Comparative Analysis Type: Analysis Objective: Compare the automated recognition system's performance with traditional manual inspection methods. Analyze the advantages and disadvantages of each approach, with a focus on potential cost savings and operational efficiency. 8 CHAPTER 4 Methodology 1. Image Resizing: To resize the original image to a desired target size, use the imresize function. 2. Grayscale Conversion: Color images can be converted to grayscale using rgb2gray. 3. Edge Detection (Canny): To find edges in a grayscale image, the Canny edge detection technique is used. The 'Canny' method and the edge function are used to accomplish this. 4. Morphological Operations: Using a disk-shaped structural element, morphological operations—specifically, closing—are done to improve edges. The imclose function is used to accomplish this. 5. Binarization (Otsu's Method): To binarize an image, automatic thresholding is used using Otsu's method. The threshold is calculated using the graythresh function, and binarization is performed using imbinarize. 6. Histogram Equalization: To improve image contrast, histogram equalization is used. For this, the histeq function is employed. 7. Radon Transform: The radon function is used to calculate the Radon Transform. To examine the image's structure, it projections it onto a range of angles. 8.Computation of the Radon Signature (Step 8): The variance of the Radon Transform values along each angle is computed as part of the Radon Signature calculation process. 9-L2-Norm Normalization: The Radon Signature is normalized to a unit vector by applying the L2-norm. 10-Data Saving: The preprocessed image and the Radon Signature plots are saved in designated output folders by the code using the imwrite function. Together, these procedures and algorithms enable the preprocessing of photos and the computation of the Radon Signatures for an image dataset. The particular dataset and the specifics of the images being processed determine the precise outcomes and traits of the Radon Signatures. 9 4.1 CHAPTER 4 Experiment and Implementation Demonstrate that you can implement the proposed approaches using OpenCVPython/ MATLAB. Describe the software/ hardware, core python packages, and how you selected parameters/ optimization of key algorithms if applicable. Use flowcharts, diagrams, algorithms and pseudo code where applicable. 4.2 Summary Summarize the methodology and highlight about expected results. 1 0 CHAPTER 5 Results and Discussion 5 Results and Discussion 5.1 Experimental Setup 5.1.1 Dataset Description Provide an overview of the dataset used for testing the recognition system. Include details such as the number of images, fabric types, and patterns. 5.1.2 Evaluation Metrics Define the evaluation metrics employed for quantitatively assessing the system's performance, including accuracy, precision, recall, and F1-score. 5.1.3 Methodology Describe the procedures followed for testing the system, including image preprocessing, feature extraction, and pattern classification. 5.2 Quantitative Results 5.2.1 Accuracy Metrics Present the accuracy results obtained from the fabric pattern recognition system for each pattern category. 5.2.2 Precision, Recall, and F1-Score Discuss precision, recall, and F1-score values for each pattern category, providing insights into the system's performance. 5.2.3 Confusion Matrix Display confusion matrices to visualize the distribution of true positives, true negatives, false positives, and false negatives for pattern classification. 5.2.4 Statistical Analysis Conduct statistical analysis, such as t-tests or ANOVA, to determine the significance of performance differences between pattern categories. 5.3 Qualitative Results 5.3.1 Pattern Classification Examples 1 1 CHAPTER 5 Include qualitative examples of fabric pattern classification to demonstrate the system's effectiveness in recognizing different patterns. 5.3.2 Visualization of Feature Extraction Visualize the features extracted by the system to provide insights into the methodology's capabilities. 5.4 Discussion 5.4.1 Performance Analysis Analyze the quantitative results, discussing the overall accuracy and metrics for each pattern category. Identify patterns with the highest and lowest recognition rates. 5.4.2 Feature Extraction Efficacy Evaluate the effectiveness of the feature extraction methods (e.g., Radon Transform, texture analysis, CNNs) and their impact on pattern recognition. 5.4.3 Integration with Human Expertise Discuss how the integration with human expertise influenced the system's performance and the user experience. 5.4.4 Comparative Analysis Compare the automated system's performance with traditional manual inspection, emphasizing any cost savings and efficiency improvements. 1 2 CHAPTER 6 Conclusion and Future Scope Conclude the project. Identify challenges relevant to your application that have not been fully resolved within this project’s scope. Propose future works to deal with these challenges. 6.1 Conclusion The successful completion of this fabric pattern recognition project represents a significant milestone in the realm of textile manufacturing and automation. The developed system, which combines computer vision, image processing, and machine learning techniques, has demonstrated its ability to accurately classify fabric patterns into categories such as plaid, checkered, horizontal/vertical stripes, dots, and irregular patterns. Through a comprehensive evaluation, both quantitatively and qualitatively, the system showcased its competence in achieving high recognition accuracy, precision, recall, and F1-scores for each pattern category. Notably, the integration of advanced feature extraction methods, including the Radon Transform and Convolutional Neural Networks (CNNs), played a pivotal role in enhancing the system's performance. The project's findings highlight the potential for a cost-efficient, objective, and efficient solution for fabric pattern recognition, replacing labor-intensive manual inspection methods. Moreover, the system's user-friendly interface and the ability to work in harmony with human expertise provide a versatile platform for various industries, including textiles, fashion, quality assurance, e-commerce, cultural preservation, research, and education. This project serves as a stepping stone toward addressing the challenges associated with manual recognition and contributes to the automation and optimization of fabric pattern recognition. While acknowledging its achievements, we also recognize the system's limitations and look forward to future work that will further refine and expand the capabilities of fabric pattern recognition technology. 6.2 Scope for Future Work While the present project has made significant strides in automating fabric pattern recognition, several challenges and opportunities for further advancement remain. The following outlines the scope for future works to address these challenges and continue the development of fabric pattern recognition technology: 1. Enhanced Pattern Classification Challenge: The recognition of intricate and nuanced fabric patterns, such as those with fine details, remains a challenge. Future Work: Future research could focus on refining the feature extraction methods and machine learning models to enhance the classification of complex patterns, possibly through the incorporation of advanced deep learning techniques or generative models. 2. Scaling and Diverse Dataset Handling 1 3 CHAPTER 6 Challenge: The system's performance in handling diverse datasets and scaling to accommodate a larger variety of fabric types and patterns could be improved. Future Work: Further research can explore techniques for dataset augmentation and finetuning to ensure the system's adaptability to a broader range of fabrics. Additionally, addressing variations in lighting conditions and image quality is a potential avenue for improvement. 3. Real-Time Recognition Challenge: Achieving real-time fabric pattern recognition to meet the demands of high-speed textile manufacturing processes remains an ongoing challenge. Future Work: Future endeavours can focus on optimizing the system's processing speed to enable real-time recognition, which would significantly impact the efficiency of fabric production. 4. Human-Machine Collaboration Challenge: The seamless integration of the automated system with human expertise and decision-making requires further exploration. Future Work: Research can delve into developing adaptive decision support systems that provide informative feedback to human inspectors, aiding in complex recognition tasks and allowing for continuous learning and collaboration. 5. Cross-Domain Applications Challenge: Extending the technology's applicability to diverse domains, such as cultural preservation and education, presents a challenge. Future Work: Future works can explore how fabric pattern recognition technology can be adapted and specialized for specific applications, providing tailored solutions for archiving historical fabric patterns or enhancing educational tools in textile studies. 6. Cost-Benefit Analysis Challenge: Comprehensive cost-benefit analyses that quantify the economic advantages of the automated system over manual inspection require further investigation. 1 4 Contribution of Each Team Member X-190929019 Detailed contribution to Mini-Project Highlight the contribution made by the team member. Analysis Include a brief analysis of the contribution. Y-180929128 Detailed contribution to Mini-Project Highlight the contribution made by the team member. Analysis Include a brief analysis of the contribution. Z-190929106 Detailed contribution to Mini-Project Highlight the contribution made by the team member. Analysis Include a brief analysis of the contribution. 1 5 REFERENCES [1] D. Cousins. “Self-driving ibex robot sprayer helps farmers safely tackle hills.” (2016), [Online]. Available: https : / / www . fwi . co . uk / livestock / grassland management/self-driving-ibex-robot-sprayer-helps-farmerssafely-tackle-hills. (accessed: 08.02.2022). [2] A. Craft. “Making it rain: Drones could be the future for cloud seeding.” (2017), [Online]. Available: https://www.foxnews.com/tech/making-it-rain-dronescould-be-the-future-for-cloud-seeding. (accessed: 08.02.2022). [3] M. Mazur. “Six ways drones are revolutionizing agriculture.” (2016), [Online]. Available: https : / / www . technologyreview . com / 2016 / 07 / 20 / 158748 / six - ways - drones - are - revolutionizing - agriculture/. (accessed: 08.02.2022). [4] Verified Market Research. “Agriculture robots market size, share, trends and opportunities.” (2018), [Online]. Available: https://www.verifiedmarketresearch. com / product / global - agriculture - robots - market - size - and forecast-to-2025/. (accessed: 08.02.2022). [5] W. Chonnaparamutt, H. Kawasaki, S. Ueki, S. Murakami, and K. Koganemaru, “Development of a timberjack-like pruning robot: Climbing experiment and fuzzy velocity control,” in 2009 ICCAS-SICE, 2009, pp. 1195–1199. [6] H. Kawasaki, S. Murakami, H. Kachi, and S. Ueki, “Novel climbing method of pruning robot,” Sep. 2008, pp. 160–163. DOI: 10.1109/SICE.2008.4654641. 1 6 Appendix Annexure to include (Optional) • Product Datasheets • Design drawings • Standard diagrams • Lengthy codes/algorithms etc 17 ANNEXURE 1 IET Learning Outcomes Student Name: Registration No.: X 190929019 Y 180929128 Z 190929106 LO M1 M2 M3 M4 M5 M7 M16 M17 Topic Name Section No. Guide’s Observation Signature of Student: Name and Signature of Guide: Date: 18 PLAGIARISM CHECK 19 General Guidelines (Delete this page when making the report submission) • Project Report to be about 25 to 40 pages. • Paper Size: A4; Left = Right = Top = Bottom Margins = 1” • Page Numbering Position: Bottom with right justified and continuous numbering from the Introduction • Use Times New Roman Font with Normal Style, paragraph justified and 1.25 line spacing • Paragraph Heading: Times New Roman Font, Bold, Font Size 14; Paragraph Matter: Times New Roman Font, Normal, Font Size 12; • Sub-paragraphs be appropriately numbered as in 1.1, 1.2, 1.3 etc; Sub-paragraph Heading: Times New Roman Font, Italics, Font Size 12; Sub-paragraph Matter: Times New Roman Font, Normal, Font Size 12; • Figure captions below Figure with continuous numbering (Ex: Fig. 1, Fig. 2, etc) • Tables captions above Table with continuous numbering (Ex: Table 1, Table 2, etc) • All references must be quoted in ascending order • Arrangement of contents as per the sequence mentioned in the format • The above guidelines should be used only as a help guide and is more or less a standard way of report writing • Report formatting should not be disturbed in any form 20