Palash Sushil Matey

Bio: Palash Sushil Matey is an academic researcher from VIT University. The author has contributed to research in topics: Multilayer perceptron & Steganography. The author has an hindex of 3, co-authored 5 publications receiving 49 citations.

Novel Method for Weed Classification in Maize Field Using Otsu and PCA Implementation

Abstract: This paper proposes two methods, oriented to crop row detection in images from agriculture fields with high weed pressure and to further distinguish between weed and crop. Firstly, for crop row detection the image processing consists of three main processes: image segmentation, double thresholding based on the 3D-Otsu's method, and crop row detection. Secondly, further classification between weed and crop, is carried out by compressing the three dimension vectors of an image to one dimension using the principal component analysis (PCA) method. Finally the combination of Otsu method and the PCA enable us to not only detect weed in crop rows but also classify this weed from crop. Hence it is better suited for the real time applications pertaining to weed detection.

Leaf recognition using contour based edge detection and SIFT algorithm

Abstract: The paper presents two advanced methods for comparative study in the field of computer vision. The first method involves the implementation of the Scalar Invariant Fourier Transform (SIFT) algorithm for the leaf recognition based on the key descriptors value. The second method involves the contour-based corner detection and classification which is done with the help of Mean Projection algorithm. The advantage of this system over the other Curvature Scale Space (CSS) systems is that there are fewer false-positive (FP) and false-negative (FN) points compared with recent standard corner detection techniques. The performance analysis of both the algorithm was done on the flavia database.

Real-time implementation of steganography in medical images using integer wavelet transform

Abstract: A method is presented to hide data within an medical image for the real time application using the Integer Wavelength Transform (IWT) technique in the transform domain with the help of steganography IWT is an image compression technique wherein the output is in the form of integers thus consuming less memory space Steganography is done intelligently such that it is difficult for an adversary to detect the existence of a hidden message in the otherwise innocuous data This united with integer wavelet transform allows high quality data hiding and image compression To transmit many such images over a network, sometimes over low-capacity phone lines to remote sites, or to store large numbers of images over a long period of time as part of the medical records for patients, the need for image compression arises to alleviate these large demands for image data storage and transmission capacity The secret information is hidden in cover image using IWT implementation which has been coded in C language to ensure easy realization in real-time applications The PSNR and execution time for the different images using IWT technique have been computed

Adaptive channel equalization using recurrent neural network under SUI channel model

Abstract: This paper investigates Neural Networks (NNs) based adaptive channel equalization for standard Stanford University Interim (SUI) channels. The NN models like Multilayer Perceptron Algorithm (MLP) and Recurrent Neural Network (RNN) are used to design adaptive equalizers. The Back Propagation (BP) and Real Time recurrent Learning (RTRL) are used for training MLP and RNN models respectively. As NNs are known for highly non-linear structure, these models are better suitable for equalization of system with high non-linearity. The performance of RNN is compared with MLP in terms of Bit Error Rate (BER). In simulation analysis, BPSK signal are transmitted through various SUI channels, which are modeled for fixed wireless applications. The simulation results illustrates that the RNN equalizer consistently outperform the MLP equalizer by giving better BER.

Equalization of Stanford University Interim channels using adaptive multilayer perceptron NN model

Abstract: This paper presents adaptive channel equalization for six standard Stanford University Interim (SUI) channels using Least Mean Square Algorithm (LMS) and Multilayer Perceptron Algorithm (MLP) models The performance analysis of the adaptive equalizers was done based on the Bit Error Rate (BER) The performance of LMS algorithm is found decent whenever there is no nonlinearity in system, whereas in presence of nonlinearity in the system, the LMS algorithm fails to perform well Under such a case, the MLP based equalizer is found to be better alternative In simulation analysis, BPSK signal are transmitted through various SUI channels The results were compared and it was found that under nonlinear conditions, MLP algorithm gives better BER in comparison to LMS

Plant Species Identification Using Computer Vision Techniques: A Systematic Literature Review

Abstract: Species knowledge is essential for protecting biodiversity. The identification of plants by conventional keys is complex, time consuming, and due to the use of specific botanical terms frustrating for non-experts. This creates a hard to overcome hurdle for novices interested in acquiring species knowledge. Today, there is an increasing interest in automating the process of species identification. The availability and ubiquity of relevant technologies, such as, digital cameras and mobile devices, the remote access to databases, new techniques in image processing and pattern recognition let the idea of automated species identification become reality. This paper is the first systematic literature review with the aim of a thorough analysis and comparison of primary studies on computer vision approaches for plant species identification. We identified 120 peer-reviewed studies, selected through a multi-stage process, published in the last 10 years (2005–2015). After a careful analysis of these studies, we describe the applied methods categorized according to the studied plant organ, and the studied features, i.e., shape, texture, color, margin, and vein structure. Furthermore, we compare methods based on classification accuracy achieved on publicly available datasets. Our results are relevant to researches in ecology as well as computer vision for their ongoing research. The systematic and concise overview will also be helpful for beginners in those research fields, as they can use the comparable analyses of applied methods as a guide in this complex activity.

An Overview of Contour Detection Approaches

Abstract: Object contour plays an important role in fields such as semantic segmentation and image classification. However, the extraction of contour is a difficult task, especially when the contour is incomplete or unclosed. In this paper, the existing contour detection approaches are reviewed and roughly divided into three categories: pixel-based, edge-based, and region-based. In addition, since the traditional contour detection approaches have achieved a high degree of sophistication, the deep convolutional neural networks (DCNNs) have good performance in image recognition, therefore, the DCNNs based contour detection approaches are also covered in this paper. Moreover, the future development of contour detection is analyzed and predicted.