Venkata Lalitha Narla

AgroDetect Smart Plant Pathology is a Web‐App‐based solution and an innovative system that utilizes image processing techniques to analyze plant diseases based on leaf images. This technology aims to revolutionize agriculture by providing a rapid and precise method of identifying and managing plant diseases, ultimately enhancing crop yield and reducing losses. The Convolutional Neural Network (CNN) model, augmented with Rectified Linear Unit (ReLU) activation functions, was employed to discern the condition of plants by analyzing leaf images. The accuracy of the performance of the designed models is evaluated using four categories of plant leaves from the PlantVillage database. The average accuracy of the designed models is around 97% and is further compared with state‐of‐the‐art works, and also design a web application using the Streamlit Python library for user interaction with minimal effort.

AgroDetect Smart Plant Pathology is a Web-App-based solution and an innovative system that utilizes image processing techniques to analyze plant diseases based on leaf images. This technology aims to revolutionize agriculture by offering a rapid and precise method of identifying and managing plant diseases, ultimately improving crop yield and reducing losses. The Convolutional Neural Network (CNN) model, augmented with Rectified Linear Unit (ReLU) activation functions, was employed to discern the condition of plants by analyzing leaf images. The accuracy of the performance of the designed models is evaluated using four categories of plant leaves from the PlantVillage database. The average accuracy of the designed models is around 97% and is further compared with state-of-art works and also designed a web application using the Streamlit Python library for user interaction with minimal effort.

One of the key forensics topics has been the detection of speech forgeries, mostly using real evidence in court. The transmission of digital speech recording data over several media exposes the data to the risk of being attacked or tampered with. Several people misuse the audio by altering that using editing software, such as Adobe, Audition CC, etc., which results in speech forgeries. So, to overcome these scenarios speech forgery detection method is deployed. A speech forgery detection method for splicing is implemented in this paper. Firstly, voiced segments are identified in the speech signal and calculated Mel Frequency Cepstral Coefficients (MFCC). These coefficients are considered as features and are stored in the database for the registered speakers. Similarly, features are calculated for each voiced segment of test signals and compared those features with the database by using dynamic time warping. This proposed method is tested on 225 original speech signals that are recorded in two different environments using two different microphones. By combining original recordings of two distinct speakers, a forged dataset of 4900 spliced speech signals is developed to test the efficacy of the developed method. An accuracy of 99.39% was attained and is superior to other existing methods.

Digital watermarking is one of the solution to protect the copyright and ownership of multimedia content. Digital audio watermarking embeds owner information as a watermark into the audio files. Synchronization is the major drawback in some of the audio watermarking techniques. In this audio watermarking algorithm based on FFT, synchronization code is included to resist the desynchronization attacks such as cropping and MP3 compression. Similarly, watermarking scheme using DWT is proposed and the performance of the two algorithms based on FFT and DWT is compared with respect to robustness and imperceptibility.

Watermarking is the process of embedding information into a signal (e.g. audio, video or pictures) in a way that is difficult to remove. In this paper, a high-capacity audio watermarking system is used to embed data and extract them using singular value decomposition (SVD). With the help of SVD-based algorithms and by using lifting wavelet transform (LWT), discrete cosine transform (DCT) and discrete wavelet transform (DWT). DCT-SVD, DWT-SVD, DWT-DCT-SVD, LWT-DCT-SVD techniques are developed. It was observed that by increasing the quantization levels the signal-to-noise ratio (SNR) value decreases exponentially which leads to distortion in the original signal. It is also observed that robustness is also increased by applying different malicious attacks like re-sampling, re-quantization, echo addition, cropping, additive white gaussian noise (AWGN), signal addition and signal subtraction to the embedded signal which doesn't disturb the original signal and the extracted image. The performance of this technique is evaluated using bit error rate (BER), cross-correlation (CC) and SNR.