Search Results for "convolution matrix"

CUTLASS is a collection of CUDA C++ template abstractions for implementing high-performance matrix-multiplication (GEMM) and related computations at all levels and scales within CUDA. It incorporates strategies for hierarchical decomposition and data movement similar to those used to implement cuBLAS and cuDNN. CUTLASS decomposes these "moving parts" into reusable, modular software components abstracted by C++ template classes. These thread-wide, warp-wide, block-wide, and device-wide...

...For the operators, both centered and upwind operators are provided, for domains of any dimension, arbitrarily spaced grids, and for any order of accuracy. The cases of 1, 2, and 3 dimensions with an evenly spaced grid are optimized with a convolution routine from NNlib.jl. Care is taken to give efficiency by avoiding unnecessary allocations, using purpose-built stencil compilers, allowing GPUs and parallelism, etc. Any operator can be concretized as an Array, a BandedMatrix or a sparse matrix.

Current filter functionality: - Simple editing options: Image cropping, resizing, rotation, Color brightness curve alignment - Histobram processing: Convolution, statistics (e. g. f_max or median analysis) - Image segmentation: The actual segmentation process as well as group weight calculation for further filtering (both functions rely on self defined custom dynamic mathematical functions) - Dynamic mathematical functions for custom and automated image filtering: General mathematical operations, using image or matrix as f(x, y), export f(x, y) as image or matrix, mapping variables on other ones and of course boolean operation for case sensitivity - A flexible variables model of dynamic mathematical function that sets no restriction on particular variables count - Sub project support for an organized total process targeting to save time using previously created editing routines instead of redoing steps each time

The blocksparse repository provides efficient GPU kernels (TensorFlow custom ops) for block-sparse matrix multiplication and convolution operations. The idea is to exploit block-level sparsity — i.e. treat matrices or weight tensors as composed of blocks, many of which may be zero or unused — to save compute and memory when sparsity patterns are structured. This is particularly useful in models like Sparse Transformers, where attention matrices or intermediate layers may adopt block-sparse patterns to scale better. ...

PhotoJShop is a JavaScript library for photo editing using the canvas and kernel convolution matrixes, aiming to reproduce the most usual filters. A demo of its capabilities can be seen in Nuophoto, a project that uses this library for all the editing. The idea is to provide a simple library to developers that will allow quick integration of photo filters to their website. After including jQuery, include photojshop.jquery.js. Once loaded you simply need to call it on a valid element. You...

Filters is a image processing library: filter sobel,convolution,morphology,vectorization,segmentation,blob,blur,histogram,susan,threshold,texture,contrast,standard deviation,canny,distance map,matrix,contour,edge,rotation,correlation,gradient anisotropic

Testlab for Image Processing: You can enter formulas like "r = r + b/2; g=g*2; b=b/2" and watch/store the Result. Alternatively cou can run a convolution matrix on the source image. Implemented in Java, fully GUIed, avg. performance.