A Novel Meta-Heuristic Jellyfish Optimizer for Detection and Recognition of Text from Complex Images
Rice Leaf Disease Detection using Convolutional Neural Network
Species Classification and Disease Identification using Image Processing and Convolutional Neural Networks
Non-Invasive Neonatal Golden Hue Detector
Comparative Analysis of Usage of Machine Learning in Image Recognition
Identification of Volcano Hotspots by using Resilient Back Propagation (RBP) Algorithm Via Satellite Images
Data Hiding in Encrypted Compressed Videos for Privacy Information Protection
Improved Video Watermarking using Discrete Cosine Transform
Contrast Enhancement based Brain Tumour MRI Image Segmentation and Detection with Low Power Consumption
Denoising of Images by Wavelets and Contourlets using Bi-Shrink Filter
Recognizing text in images presents significant challenges, particularly in complex backgrounds. This technology is essential for aiding visually impaired individuals and interpreting semantic content. This survey examines various techniques developed in recent years for handling text recognition in complex images. The paper provides an analysis of related works and evaluates the performance of these recognition methods. Although image complexity is not easily defined, it can be described through parameters such as background details, noise levels, lighting conditions, textures, and fonts. Additionally, this survey discusses several benchmark datasets used in the reviewed studies. By reviewing these works, challenges in this area can be identified and compared. Complex background images limit the accuracy achieved. To improve accuracy, convolutional neural networks (CNN) are employed. The proposed method comprises three parts: the text is detected from a complex background, the text is extracted from the image using Tesseract, and all identified words are saved in a text file. An audio file is then created from the text. The proposed system reads the text from the image with the aim of providing assistance to visually impaired individuals.
Rice is one of the most widely cultivated crops in India, playing a crucial role in the country's agricultural economy. However, rice production is frequently threatened by diseases, such as Bacterial Leaf Blight, Brown Spots, and Sheath Blight, which can significantly reduce yield. This study focuses on developing a deep learning based rice leaf disease detection system using Convolutional Neural Networks (CNN). A dataset containing images of rice leaves, categorized into different disease types and healthy leaves, was used to train the model. By applying advanced image processing and deep learning techniques, the system can accurately identify and classify diseases. A user-friendly web application allows farmers to upload images and receive real-time diagnostic feedback, empowering them to implement timely corrective measures. This system offers a scalable and cost-effective solution for enhancing crop management, supporting food security, and promoting sustainable agriculture in India.
The "Integrated Plant Disease Detection System (IPDDS)" presents a comprehensive approach to plant health assessment, combining image processing and convolutional neural networks (CNN) for species classification and disease identification. In the initial phase, input images undergo preprocessing to remove noise and enhance clarity, followed by segmentation using k-means clustering for effective region identification. A CNN classifier then utilizes deep learning techniques to categorize the plants into distinct species such as apple, core, grape, pepper bell, potato, or tomato. Subsequently, the system employs an improved CNN architecture adapted for disease classification, distinguishing various diseases affecting each plant species. For instance, diseases like Black Rot, Scab, and Cedar Rust are identified in apples, while Common Rust, Northern Blight, and Cercospora are detected in corn. This methodology enhances accuracy and reliability in disease detection, enabling timely interventions to mitigate crop losses. Furthermore, the system suggests suitable fertilizers based on disease diagnosis, facilitating targeted disease management strategies. This integrated approach offers a promising solution for effective plant disease detection, contributing to sustainable agriculture and food security.
Jaundice is a common neonatal condition that causes yellowing of the eyes and skin due to elevated pigment levels. Although several underlying disorders may contribute, the condition typically results from the immature liver function in newborns., Approximately 60% of babies experienced jaundice, leading to yellowing of the skin and eyes over the last five years. Due to improved screening and management procedures, the number of severe jaundice cases requiring treatment has decreased by about 20%. This condition generally resolves as the baby's liver matures and becomes more efficient at metabolizing the pigment. The "Golden Hue Detector," a non-invasive device specifically designed to detect neonatal jaundice, was developed and validated in this study. Early and accurate diagnosis of jaundice is crucial to prevent complications such as kernicterus. Traditional detection methods rely on blood tests, which can be stressful for infants and resource-intensive for healthcare providers. These methods typically require multiple blood draws, increasing the risk of infection and causing significant discomfort for newborns.
This paper provides an overview of the advancements and challenges in image recognition technology that utilizes machine learning. Image recognition involves identifying and categorizing objects in images, with the primary goal of determining the best algorithm for accurately classifying these objects. The paper outlines essential processes in image recognition, such as data acquisition, preprocessing, feature extraction, and classifier design, highlighting the importance of feature extraction and selection for improving accuracy. A comparison of ML and DL techniques, as well as algorithms used in image recognition, is discussed, along with the classification of images using machine learning. The aim is to identify the processes, models, and algorithms most suitable for classifying and recognizing detected objects. Furthermore, the research explores the evolution of image recognition, particularly through the combination of ML and DL techniques. By examining the broad applications of ML in this field, the paper aims to offer insights that can enhance the development of machine learning applications in image recognition, thereby improving the effectiveness of image processing technologies by identifying the best algorithms and processes.
