Category: ai
-
Radiant Vision: Unveiling True Colors with Retinex Image Enhancement
Retinex is an image enhancement technique inspired by human vision. It separates an image into illumination and reflectance components to correct uneven lighting. It is based on the Retinex theory of color vision, which posits that the perceived color of an object is determined by its reflectance properties, rather than solely by the illumination it Read.
-
What Are try and catch in C++?
Recently, I traveled to Nagoya on a business trip to deliver our camera-calibration software and train the client’s team. Once I arrived, I discovered that the application would occasionally terminate unexpectedly because an internal optimization routine diverged—usually triggered by low-quality images. Although I consider such failures a normal part of calibration, the abrupt crashes understandably Read.
-
Python Built-in Container Types
Python includes ten core container types organised into three categories: sequences, sets, and mappings. Each container holds references to other objects and supports membership tests (in). Sequences (7 Types) Sequences are ordered and indexable collections. Some are mutable; others are not. 1. list Mutable, ordered collection of arbitrary objects Allows duplicates Syntax: fruits = ['apple', Read.
-
Machine learning-ML
Machine learning (ML) is like teaching a computer to learn from examples instead of programming it to follow specific rules. Instead of saying, "If the image has two round shapes and a stem, then it’s an apple," you show the computer hundreds or thousands of images of apples and let it figure out the patterns Read.
-
Vectors insert function in C++
Vectors insert: The below code is an example of a vector insert function. It may help the beginners. #include <iostream> #include <vector> int main() { std::vector<int> v = {1, 2, 4, 5}; // Insert a single element at position 2 v.insert(v.begin() + 2, 3); // v = {1, 2, 3, 4, 5} // Insert multiple Read.
-
A single‐header C++ neural-network!
I wrote a code sample of a Neural network written in C++ that utilises only three popular libraries. This code has the following abilities: 1. Specify any number of inputs/outputs 2. Add as many hidden layers as you like 3. Pick Sigmoid or ReLU activations 4. Do forward passes and simple SGD training via back-propagation Read.
-
How to get depth image from stereo left and right image
If you have two images from a stereo camera, then you can convert them to a depth image using the function below. cv::Mat getDepthImage( cv::Mat& left_row, cv::Mat& right_row) { // for creating path std::string str = workingFolder + str_cam + "\\"; std::string name = stereo_name + cam_no; std::string fpath = str + name + ".xml"; Read.
-
cv::StereoSGBM::create(…) in OpenCV
The StereoSGBM (Semi-Global Block Matching) algorithm computes a disparity map from a pair of stereo images. This disparity map represents the pixel shift between matching points in the left and right images, which in turn relates to depth. The cv::StereoSGBM::create(…) function initializes the algorithm with the following tuning parameters: Parameter Explanations: 1. minDisparity (int) o Read.
-
Physics-Informed Neural Networks
Physics-Informed Neural Networks (PINNs) are an innovative class of deep neural network models that directly embed the laws of physics—typically formulated as partial or ordinary differential equations (PDEs or ODEs)—into the training process. Instead of relying solely on large volumes of data, PINNs augment the traditional data-driven loss with additional terms that penalise deviations from Read.
-
Gradient Descent- an optimization technique
Gradient descent is the quiet workhorse behind many of today’s smartest algorithms. At its heart, it’s a simple process of “feel, step, repeat” that guides a model from cluelessness toward competence—much like a hiker fumbling down a foggy mountainside until they find the valley floor. In this essay, we’ll explore what gradient descent is, why Read.
Khairuzzaman Mamun 