Biswajit Biswas's Projects
This code impliments a multi-focus color image fusion method based on focus region detection with edge preseve using bilateral filter and guided filter in discrete wavelet transform (2DWT). Firstly, a novel focus region detection method is estimated, which uses guided filter to refine the rough focus maps obtained by bilateral filter and difference operator. Then, An initial decision map is got via the pixel-wise maximum rule, and optimized to generate final decision map by using guided filter again. Finally, the fused image is obtained by the pixel-wise weighted-averaging rule with the final decision map via inverse transform.
I have implemented a particle effect in NVIDIA Cg Tool Kit using OpenMP to render the effect. This demo uses the vertex and fragment shader pipeline and the device processor (GPU) to perform the n-particles simulation.
This code is designed to solve the heat equation in a 2D plate with CUDA-Opengl. After solution, graphical simulation appears to show that how the heat diffuses throughout the medium within time interval selected in the code. Solving the 2 Dimensional Heat conduction equation in the generalized form, we used FEM technique.
Using Tensorflow and transfer learning, easily make a labeled image classifier with convolutional neural network
AKAZE local features matching
Automatic Polyps Segmentation using Deep CNNs
Automatic SAR Segmentation using Deep CNNs
Binary Image classification using ImageDataGenerator in Keras
C++ code practice intel c compiler
Automatic Camera calibration using OpenCV an ViPS
This algorithm performs automated detection, measurement, and levelling of co-cultured cells. Each cell is identified and characterized by probabilistic hesitant fuzzy set with several local descriptors (entropy, fuzzy index, fuzzy connectivity,etc. ), which are used to define an optimal threshold to segment individual cells.
This code is capable of simulating of flows through materials with complex, porous structure. In this way, the whole process of importing files containing the porous geometry, generating the computational mesh and running CFD simulation can be performed in MATLAB
The CImg Library is a small and open-source C++ toolkit for image processing
we have just coded to implement parallel Newton’s recursive algorithm. To accomplish this, we first partition the problem and fitted each partition into master-slave by MPI mode. First, we need to have: (a) The number of interpolation points. (b) A grid of points at which the interpolation is to be exact. (c) An array containing the function we wish to interpolate evaluated at the interpolating grid. (d) An array to store the Newton differencing coefficients
Run a CUDA kernel that writes image data to a GL buffer or texture image. A live display via CUDA Graphics Interop mode on Ubuntu and CUDA-11.0 Driver.
Run a CUDA kernel that writes image data to a GL buffer or texture image. A live display via CUDA Graphics Interop mode on Ubuntu and CUDA-11.0 Driver.
Color image quantization is a process of selecting a set of colors to display an image with some representative colors, this without noticeable perceived difference. This algorithm deals with the problem of the quantization errors evaluation in taking into account a probabilistic measure on color space. Developed a imageJ plugin to minimize the possible degradationscolor profile.
This gives a list of tools that will improve different computer vision projects . A number of concepts were developed, including approaches to solve computer vision and image processing algorithms.
C++ Codes for different Computer Vision algorithms
contour detection for color image using Visp c++ library
Convex hull reconstruction and partitioning from 3D RGBD scanned cloud data by pcl and cgal
Convex hull construction from Cloud Points using pcl and cgal
Building a real-time computer vision algorithms using Tensorflow and OpenCV library in Python
Several deep learning applications have design using Python on GPU on Linux
deep convolutional autoencoder for image denoising
To display very high resolution video directly with OpenGL and OpenCV on Qt5
A website showcasing 3rd party software designed for elementary OS 0.3 Freya
Deep Image Patch (DIP) correction method uses to reconstruct a clean image by deep convolutional neural networks (DCNNs) traing on noisy images with different noise model and ground truth.
The fastai deep learning library
A deep learning method, ingraft-U-Net, is proposed to segment polyps using Melanoma Skin frames. Ingraft-U-Net is a modified version of UNet, which comprises three stages, including the preprocessing, encoder, and decoder stages.