Installing OpenCV 4 in Mac Os

Installing OpenCV 4 in Mac Os Install Anaconda Create a new environment Open the new environment in the terminal Run: conda install -c conda-forge/label/main opencv conda install notebook Jupyter notebook In the notebook import cv2 Check the version with cv2.__version__ You should see something like 4.1.1 Problems when showing an image or video If when…

What I Learned by Teaching Computer Vision in an AI Bootcamp

The opportunity In July 2019 I was in a meeting with the CEO of akademy.ai where I received an offer to teach computer vision at their AI Engineer Bootcamp here in Barcelona. The bootcamp consists of a fully immersive 10 weeks program to launch your career in Artificial Intelligence. It was a very interesting offer…

My research about Jobs in Computer Vision and how you might get one

During the #100DaysOfMLCode challenge I was going through the Computer Vision Udacity nanodegree and I had the opportunity to learn a lot while doing it, I complemented the nanodegree with other computer vision courses and books and while doing it I also did some research about computer vision jobs, and this is what I came…

Computer vision algorithms

ORB Algorithm: Oriented Fast and Rotated Brief, creates feature vectors from detected keypoints and is invariant to rotations, changes in illumination, and noise. HOG Algorithm: Histogram of Oriented Gradients works by creating histograms of the distribution of gradient orientations in an image and then normalizing them in a very special way. This special normalization is…

Computer Vision: Types of Features and Image Segmentation

Types of Features and Image Segmentation Types of features: Edges, Corners and Blobs. Corner Detector: Intersection of 2 edges, can be calculated with Sobel operators (Sobel x and Sobel y) Gx and Gy (G for Gradient) Dilation (add pixels to the boundaries of an object) and erosion (removes pixels along object boundaries) can be combined to fill in gaps in…

Localization

Robot localization in essence is based in two main steps: Sense and Move It will start with a initial belief (or prior) of maximum confusion where the probability distribution will be uniform (flat, which means it has the same value everywhere) Then it will start cycling through sensor measurements (Sense) and movements (Move) When the robot moves…

Motion in Computer Vision

Motion in Computer Vision Motion can be tracked with a 2D motion vector. A vector has a direction and magnitude which will determine the direction and amount of movement between one frame and the next one. First we will have to defive special points to track like for example intersections or corners once we localize…

Image Captioning Project

In this project I will train a network with the COCO Dataset (Common Objects in Context). This dataset contains images and a set of 5 different captions per image. I will train a CNN-RNN model by feeding it with the image and captions so the network will learn to generate captions given an image. Once trained…

Attention Mechanisms

Allows the network to pay more attention to the most important parts, so in an image it will find the most important pixels or in language transltation it will know what words to focus on and the order it has to process them. The inputs are passed in to an encoder, the encoder will generate…

YOLO (You Only Look Once)

YOLO (You Only Look Once) It is a real time object detection algorithm which prioritizes speed and recognition. YOLO will output the probabilities of the classes as well as the bounding box parameters (x,y,w,h) where x and y determines the coordinates of the center of the box and w and h are the width…