December 2024 – May 2025 | Bangalore, India
August 2024 – Present | Hyderabad, India
September 2024 - Present | Hyderabad, India
January 2024 - Present | Hyderabad, India
March 2023 - January 2024 | Hyderabad, India
July 2023 - October 2023 | Hyderabad, India
June 2022 – May 2023 | Hyderabad, India
July 2022 - July 2023 | Hyderabad, India
May 2021 - June 2021 | Singapore
Depression is a common mental health disease that requires early identification for successful treatment. Through voice analysis, this study aims to create an automated framework using convolutional neural networks (CNN) for the early detection and prediction of depression. The study looks at different types of depression, causes of sadness, ways to prevent it, and the ability of automated speech analysis to identify depressive symptoms. A thorough examination of the optimised multi-channel weighted speech classification (OMCWSC) system is also provided, emphasising its efficacy in foretelling adolescent depression. The results highlight the importance of acoustic speech characteristics and the promising potential of deep learning methods in depression analysis.
Sesame is a major oilseed crop in India. Among the major yield limiting factors, biotic stress plays the major role in yield reduction which leads to decrease in cultivable areas. Sesame phyllody disease is a very serious disease in most sesame growing regions and causes up to 100% yield loss and thus its management gained national importance. Phyllody caused by phytoplasmas where the entire inflorescence is replaced by a growth consisting of short, twisted leaves closely arranged on a stem with very short internodes and affected plants remain partially or completely sterile, resulting in total loss in yield. The disease spreads in nature by different leafhopper species, among them Orosius orientalis was reported as the major vector. No resistant germplasm has been identified till date and no precise integrated management approaches are available to combat the disease. In this study we are attempting to predict the disease using computer vision and deep learning models. The study presents a detailed exploration of Mobile NetV2 for addressing the disease detection problem, emphasizing the optimization of accuracy and efficiency. Transfer learning and hyperparameter tuning were employed to enhance model performance, and our findings show significant improvements in key metrics.