Search Results for "per linux"

VGGT is a transformer-based framework aimed at unifying classic visual geometry tasks—such as depth estimation, camera pose recovery, point tracking, and correspondence—under a single model. Rather than training separate networks per task, it shares an encoder and leverages geometric heads/decoders to infer structure and motion from images or short clips. The design emphasizes consistent geometric reasoning: outputs from one head (e.g., correspondences or tracks) reinforce others (e.g., pose...

This repository implements Class Activation Mapping (CAM), a technique to expose the implicit attention of convolutional neural networks by generating heatmaps that highlight the most discriminative image regions influencing a network’s class prediction. The method involves modifying a CNN model slightly (e.g., using global average pooling before the final layer) to produce a weighted combination of feature maps as the class activation map. Integration with existing CNNs (with light...

butteraugli is a perceptual similarity metric designed to estimate how noticeable differences between two images will be to the human eye. Instead of simple pixel math, it models aspects of human vision—color sensitivity, spatial masking, and contrast perception—to highlight differences that viewers actually see. The core tool outputs a single “distance” score along with per-pixel or per-region maps that show where artifacts are most objectionable. These maps make it practical to tune...

R-FCN (“Region-based Fully Convolutional Networks”) is an object detection framework that makes almost all computation fully convolutional and shared across the image, unlike prior region-based approaches (e.g. Faster R-CNN) which run per-region sub-networks. The repository provides an implementation (in Python) supporting end-to-end training and inference of R-FCN models on standard datasets. The authors propose position-sensitive score maps to reconcile the need for translation variance...

Structured Edge Detection (Edges) is a MATLAB toolbox implementing the structured forests method for fast and accurate edge detection (up to ~60 fps in many settings). The toolbox also includes the Edge Boxes object proposal method, fast superpixel generation, and utilities for training, evaluation, and integration with vision pipelines. High performance (frames per second performance depending on settings). Integration with MATLAB and compatibility with external vision pipelines. Fast edge...

Video processing and computer vision library for GNU/Linux offering interfaces to do image- and video-I/O with ImageMagick/Magick++, Xine, firewire digital camera (DC1394), and video for linux (V4L2). Note that this version of HornetsEye is deprecated. HornetsEye now is released as multiple packages on RubyGems.org. The source code is available on Github.com (see https://wedesoft.github.io/hornetseye-doc/ for more information).