Recent News

  • Faculty and Students Leads to Patent Publication for Intelligent Shelf Management System July 22, 2024

    In a groundbreaking collaboration, Dr Banee Bandana Das, Assistant Professor in the Department of Computer Science and Engineering, and Dr Saswat Kumar Ram, Assistant Professor in the Department of Electronics and Communication Engineering, have joined forces with Btech-CSE students Mr Rohit Kumar Jupalle, Mr Dinesh Sai Sandeep Desu, and Mr Nikhil Kethavath to develop and patent an innovative invention.,” The team’s invention, titled “A SYSTEM FOR VISION-BASED INTELLIGENT SHELF MANAGEMENT SYSTEM AND A METHOD THEREOF,” has been officially filed and published with Application Number 202441039394 in the Patent Office Journal. This invention showcases the academic excellence and collaborative spirit within the institution, as faculty members and students work together to push the boundaries of technology and create solutions with real-world impact.
    This significant achievement not only highlights the creativity and dedication of the individuals involved but also underscores the institution’s commitment to fostering a culture of innovation and research. The publication of this invention paves the way for further exploration and development in the field of intelligent shelf management systems, demonstrating the potential for transformative contributions to the industry.

    This research offers the best solution to improve the business in the retail realm, maintaining On- Shelf Availability (OSA) is vital for customer satisfaction and profitability. Traditional OSA methods face accuracy challenges, prompting a shift to deep learning models like YOLO and CNN. However, data quality remains a hurdle. This research introduces OSA, a novel semi-supervised approach merging ’semi-supervised learning’ and ’on-shelf availability’ with YOLO. It reduces human effort and computation time, focusing on efficient empty-shelf detection. Implementing a Vision-Based Intelligent Shelf Management System empowers retailers with real-time insights, revolutionizing decision-making. The model is optimized for diverse devices and provides practical solutions for efficient retail operations. Balancing model complexity, size, latency, and accuracy, the research paves the way for an advanced, data-driven shelf management approach, contributing to improved shopping experiences and business profitability

    Practical Implementation and the Social Implications Associated
    1. The present invention is a time-saving method in maintaining the stocks.
    2. The use Vision-Based Intelligent Shelf-Management System provide a well alternative in reducing the labor efforts.
    3. The system will help in terms of self-management system using machine learning techniques to optimize restocking decisions.
    The present invention can be used in shopping malls and business areas for enhancing customer experiences and business and few application areas are:
    • Smart City and smart Village
    This technique and system can reduce the human efforts in identifying vacant slots for items in business areas and provides necessary inputs to fill the same within a time frame.
    • Automobile Industry
    The system can be easily integrated with the showrooms to identify the empty spaces and inform to get it fill with products.

    SRM AP Faculties and UROP Students

    Future research plan

    In the future, different deep learning and machine learning methods can be merged to explore better performance in identifying overlapping objects.

    Continue reading →
  • Dr Sanjay Kumar and Team Publish Digital Image Security Invention in the Patent Office Journal July 11, 2024

    In a recent development, Dr Sanjay Kumar, Assistant Professor in the Department of Computer Science and Engineering, along with a team comprising N Ujwala, J Yashwanth Sri Ram, J Lakshmi Prasanna Kumar, and K Heman Rai Choudhary, have successfully filed and published an innovative invention in the Patent Office Journal titled “SYSTEM AND METHOD FOR DISCRETE WAVELET TRANSFORM (DWT)-BASED SECURE IMAGE WATERMARKING” with the application number “202441045387”.

    This invention holds great promise in the field of digital image security and watermarking. The patented system and method are based on the discrete wavelet transform (DWT), a widely used signal processing technique, to embed secure watermarks into digital images. The application of DWT ensures that the watermarking process is robust and secure, making it suitable for a wide range of applications where image integrity and authenticity need to be ensured.

    Dr Sanjay Kumar’s expertise as an Assistant Professor, coupled with the team’s skills and dedication, has resulted in the successful development and patenting of this cutting-edge technology. The “SYSTEM AND METHOD FOR DISCRETE WAVELET TRANSFORM (DWT)-BASED SECURE IMAGE WATERMARKING” promises to be a significant addition to the field of digital image security and watermarking, offering enhanced protection against unauthorised tampering and misuse of digital images. It is expected that this invention will garner attention from industry professionals, researchers, and policymakers, paving the way for its integration into diverse digital imaging systems.


    Our research introduces a robust image watermarking technique that combines Discrete Wavelet Transform (DWT) and chaotic map-based encryption. The method analyzes high-frequency sub-bands derived from DWT applied to the blue channel of an RGB image, selecting the block with the highest energy for embedding a grayscale watermark encrypted with the Henon Map. The alpha blending technique is used to integrate the encrypted watermark, ensuring both imperceptibility and robustness. The method achieves an average PSNR of 43.7211 dB and SSIM of 0.9950. The watermark can be extracted by analyzing patterns in the high-frequency component, even after various attacks, using inverse DWT and Henon Map for decryption.

    Explanation of the Patent in Layperson’s Terms
    Our research focuses on protecting digital images by embedding a hidden watermark that is hard to remove. We use a mathematical method called the Discrete Wavelet Transform (DWT) to break down an image into different parts and find the best place to hide the watermark. The watermark is further secured by encrypting it with a technique called the Henon Map. Our method ensures that the watermark remains invisible to the naked eye while being resistant to tampering. This means the watermark can be detected and recovered even if the image is altered.

    Practical Implementation and the Social Implications

    The primary application of our research is in protecting the ownership and integrity of digital images. This technique can be used by photographers, artists, and digital content creators to ensure their work is not copied or altered without permission. By embedding a secure, invisible watermark, they can prove ownership and detect unauthorised use. Additionally, this method can be applied in sensitive fields such as medical imaging and legal documents where tamper detection is crucial.


    This research was conducted by the Visual Information Processing Lab at the Department of Computer Science and Engineering, SRM University AP, Guntur, India. The team comprised Nadella Ujwala, Sanjay Kumar, Jayyavarapu Yaswanth Sri Ram, Jetti Lakshmi Prasanna Kumar, and Katragadda Heman Rai Choudhary.

    Future Research Plans

    Our future research will focus on enhancing the watermarking technique’s robustness against more sophisticated attacks, exploring real-time applications in video watermarking, and developing methods to embed multiple watermarks in a single image. We also aim to reduce the computational complexity to make the algorithm more efficient for practical applications.

    Continue reading →