International Research Journal of Engineering and Technology (IRJET) Volume: 11 Issue: 04 | Apr 2024
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
SURVEY PAPER ON INDOOR OBJECT DETECTION AND VOICE FEEDBACK SYSTEM Beeta Narayan1, Aiswariya Binu2, Anagha.P3, Anandu Ajithkumar4, A V Guruprasad5 Department of Computer Science and Engineering,Sree Narayana Gurukulam College Of Enginnering,Kadayiruppu,Ernakulam,Kerala,India ---------------------------------------------------------------------***--------------------------------------------------------------------Abstract—This paper surveys recent advances in object detection technologies with a special focus on applications designed to assist visually impaired people. We review a range of methodologies, including the You Only Look Once (YOLO) algorithm, convolutional neural networks (CNNs), and their integration with voice feedback mechanisms. A key study highlighted is the "Object Detection with Voice Feedback" project, which uses YOLO v3 for real-time object detection and provides auditory feedback through text-tospeech technology. This survey aims to synthesize the current landscape, compare different approaches, and identify future directions for research in assistive technologies for the visually impaired.
refinement of existing technologies but also the exploration of new frontiers in assistive technology. This includes improving the accuracy and speed of object detection algorithms, making devices more user-friendly and adaptable to various environments, and fostering interdisciplinary collaboration to address the multifaceted challenges faced by the visually impaired community.
II. METHODOLOGY The review methodology for this survey on AI-based assistive technologies for the visually impaired involves two main phases: literature collection and analysis. Initially, a thorough search of academic databases and journals is performed using specific keywords to gather relevant studies. Publications are selected based on strict criteria to ensure they are both pertinent and of high quality. This step aims to filter out irrelevant or lowquality papers, focusing only on those that significantly contribute to the field.
Keywords: Object detection, YOLO, Deep neural network, Tensor flow, OpenCV, Python, Raspberry Pi3b+, Google Text To Speech.
I. INTRODUCTION The advent of Artificial Intelligence (AI) and Machine Learning (ML) technologies has ushered in a new era of possibilities for enhancing the lives of individuals with visual impairments. The development and refinement of assistive devices powered by sophisticated AI algorithms, particularly in the domain of object detection and identification, have shown immense potential in bridging the gap between the capabilities of the visually impaired and the demands of navigating through a visually oriented world. These technologies leverage the power of Convolutional Neural Networks (CNNs) and innovative models like You Only Look Once (YOLO) to provide realtime information about the surrounding environment, thereby empowering users with crucial data to make informed decisions and interact more freely and safely with their surroundings.
The second phase involves a detailed examination and synthesis of the selected studies. Key information is extracted, including the types of AI technologies used, their application in assistive tools for the visually impaired, and the outcomes of these applications. This stage also assesses the quality and reliability of the studies to ensure the findings are robust and credible. The methodology is designed to provide a comprehensive overview of current trends and identify potential areas for future research, ensuring a concise yet thorough exploration of AI's role in enhancing the lives of individuals with visual impairments. Object detection technologies have revolutionized the way computers interpret visual data, enabling machines to identify and locate objects within images or videos. This field, a crucial aspect of artificial intelligence and computer vision, has evolved significantly from traditional methods to advanced deep learning approaches. Traditional techniques relied on manually coded algorithms to recognize objects, which were effective yet limited in complexity and adaptability. The introduction of deep learning, particularly Convolutional Neural Networks (CNNs), marked a significant
This burgeoning field stands at the intersection of technology and accessibility, aiming not just to compensate for visual impairments but to revolutionize the way affected individuals engage with their environment. By translating visual data into audible or tactile feedback, these AI-driven systems offer a semblance of sight, thus enhancing the quality of life and independence of their users. The ongoing research and development in this area promise not only the
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 82