KRK. I. Ramachandran

K. I. Ramachandran

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


K. I. Ramachandran has a total of 36 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Rapid Method to Rank Sound Sources in a Reverberant Environment

    by Lakshmikanthan Chinnasamy, K. I. Ramachandran

    Sound generated by an engine is one of the major contributors to an automobile’s overall sound pressure level (SPL). During its development, the engine’s SPL is monitored frequently for compliance. If the SPL is found to be exceeded, sources are identified for modification. As the timeline is critical, a rapid methodology is required to reduce it. The proposed methodology employs dynamic mode decomposition (DMD), a data-driven approach for source ranking. DMD detects changes in a variable between two successive time intervals and quantifies its rate of change. Based on the physics involved, a new interpretation for DMD’s decay factor is introduced to capture the sound signal’s nature. DMD and Short-Time Fourier Transform (STFT) are compared in this study. The STFT-based systems employ a large array of microphones to rank the sources. In this study, a single microphone and DMD are used to rank the sources. DMD utilizes far less data than the conventional STFT for source identification and ranking. Though DMD requires more memory and time than STFT, the overall time to complete the ranking of sources is very short. The results are compared with calculated values and show a negligible discrepancy between them.

  • Open Access

    Attention‑Enhanced Convolutional Neural Network for Music Genre Classification

    by M Pushparajan, Kt Sreekumar, Ki Ramachandran, C Santhosh Kumar

    The complex interplay of harmonic, rhythmic, and timbral elements within musical signals poses a challenge for accurate music genre classification. Conventional approaches have limitations in effectively capturing both localized features and long‑range contextual relationships within the audio data. To overcome this, we introduce an attention‑enhanced convolutional neural network (AECNN) architecture that employs convolutional layers to efficiently extract localized features and an attention mechanism to capture long‑range temporal dependencies within the input audio segment. This integrated approach empowers AECNN to comprehensively learn both local features and segment‑level contextual relationships, resulting in enhanced accuracy and robustness in classifying diverse music genres. The performance of the proposed model is evaluated on two datasets, the widely used GTZAN dataset and the novel IMG‑Amrita dataset, encompassing 12 distinct Indian music genres. Experiments conducted with varying audio sample durations (30, 15, 10, 6, and 3 s) demonstrate the impact of sample length on classification accuracy. Notably, AECNN achieves peak accuracies of 96.02% on the GTZAN dataset and 98.94% on the IMG‑Amrita dataset with 3‑s samples, underscoring its ability to capture intricate musical nuances with limited audio data. These findings highlight the significant potential of the AECNN for robust and versatile performance across a wide range of music classification tasks.

  • Open Access

    Improving Epileptic Seizure Prediction Using ECG From Wearable Sensors Through Spectral Whitening and Cepstral Processing

    by Pooja Muralidharan, C. Santhosh Kumar, Ramshekhar N. Menon, Haizhou Li, K. I. Ramachandran, Ravi Sankaran


  • Open Access

    Transformer-Based Framework for GPS and Timestamp Extraction from Dashcam Videos

    by Selvamani Chandrasekaran, Ramachandran K I, Adarsh S, Binoy B Nair, Aravinth J

    Dashboard cameras are becoming increasingly prevalent in vehicles, leading to a significant demand for reliable methods to extract essential metadata, such as timestamps, geolocation, and speed, from recorded footage. This metadata is essential for contextualizing events captured by these cameras, facilitating tasks such as accident reconstruction, security monitoring, and forensic analysis. However, many modern low-cost dashboard cameras overlay text directly onto video rather than logging metadata separately. This study aims to develop a solution that is independent of specific devices for extracting metadata from dash- board camera footage. Optical Character Recognition (OCR) technology presents a promising avenue due to its hardware-agnostic capabilities. The research investigates the application of pre-trained OCR models, including Tesseract, KerasOCR, and EasyOCR, to extract overlaid data from images and videos cap- tured by dashboard cameras, while also assessing the limitations and potential failure scenarios of these models. The findings indicate that a combination of traditional OCR algorithms and preprocessing techniques achieves an average Character Recognition Rate (CRR) of 50.6% and a Character Error Rate (CER) of 49.4%. To address the shortcomings of traditional OCR methods, a novel Transformer-based Optical Character Recognition (TrOCR) approach is pro- posed. Extensive training and validation of the TrOCR model, utilizing a mixed 1dataset of real and synthetic footage, significantly enhance character recognition accuracy to 84%, with a corresponding reduction in character error rate to 16%. Furthermore, the incorporation of post-processing techniques results in excep- tional accuracy of 97% and a negligible character error rate of only 0.08%. This research introduces a device-independent approach to extract metadata from dashboard camera footage, demonstrating the efficacy of the TrOCR method and underscoring the benefits of post-processing techniques for live data extraction.

  • Open Access

    A Comparative Study of Image Representation for Roller Bearing Fault Diagnosis Using Pretrained Networks

    by M. Arun Balaji, S. Naveen Venkatesh, V. Sugumaran, K. I. Ramachandran

    Roller bearings are critical components in many types of machinery, and their failure may cause significant downtime and maintenance costs. Fault diagnosis of roller bearings is thus crucial for detecting potential problems before they cause catastrophic failure and for planning maintenance and repair operations ahead of time. Early detection of roller bearing failures can help to minimize costly machine downtime and save maintenance costs. This study uses the help of deep learning models for roller bearing fault diagnosis, which can help to minimize machinery downtime and maintenance costs. The study utilizes 12 deep learning modules, and they were evaluated using various image generation methods such as vibration plot, radar plot, polar plot, Hilbert–Huang transforms, spectrogram, and scalogram. From the experimental findings, the ResNet18 model has achieved a 100.00% accuracy when the spectrogram image generation method was employed. The findings highlight the importance of selecting and optimizing deep learning models for a specific maintenance application and contribute valuable insights for researchers and practitioners in reliability engineering.


  • Fusion of Ultrasonic and RP Lidar 360 sensors using ANFIS for mobile robot navigation

    by R Ramkumar, S Adarsh, K I Ramachandran

    Multi-sensor data fusion using ANFIS for mobile robot navigation is implemented in this paper. The proposed system fuses the sensor data from ultrasonic and Lidar sensors for an improved environment perception. Using two ultrasonic sensors and one RP Lidar 360 o sensor a mobile robot test bed is created. The mobile robot is allowed to navigate through a mazelike path for collecting the real-time distance data. An ANFIS system based on these collected distance data is designed. ANFIS generates input and output membership functions by studying nature of the loaded data. Output from ANFIS is the fused distance of ultrasonic and RP Lidar sensors. Detailed statistical analysis using RMSE is done for various type of membership functions and various number of membership functions in ANFIS. Now a fuzzy logic controller consists of FIS model system decides actions to navigate autonomous robot based on the fused distance output from ANFIS.


  • Open Access

    Fault Diagnosis of Bearings using Vibration Signals and Wavelets

    by V. Dhanush Abhijit, V. Sugumaran, K. I. Ramachandran

    Fault Diagnosis of Bearings using Vibration Signals and Wavelets


  • Open Access

    Three dimensional design, simulation and optimization of a novel, universal diabetic foot offloading orthosis

    by Chand Sukumar, K I Ramachandran

    Leg amputation is a major consequence of aggregated foot ulceration in diabetic patients. A common sense based treatment approach for diabetic foot ulceration is foot offloading where the patient is required to wear a foot offloading orthosis during the entire treatment course. Removable walker is an excellent foot offloading modality compared to the golden standard solution - total contact cast and felt padding. Commercially available foot offloaders are generally customized with huge cost and less patient compliance. This work suggests an optimized 3D model of a new type light weight removable foot offloading orthosis for diabetic patients. The device has simple adjustable features which make this suitable for wide range of patients with weight of 35 to 74 kg and height of 137 to 180 cm. Foot plate of this orthosis is unisexual, with a size adjustability of (US size) 6 to 10. Materials like Aluminum alloy 6061-T6, Acrylonitrile Butadiene Styrene (ABS) and Polyurethane acted as the key player in reducing weight of the device to 0.804 kg. Static analysis of this device indicated that maximum stress developed in this device under a load of 1000 N is only 37.8 MPa, with a small deflection of 0.150 cm and factor of safety of 3.28, keeping the safety limits, whereas dynamic analysis results assures the load bearing capacity of this device. Thus, the proposed device can be safely used as an orthosis for offloading diabetic ulcerated foot.