Computer Vision News - August 2016

Let’s see the function for finding the line in the image with the Hough transform. The function first converts the image from RGB to grayscale. Next, it extracts rough edges using the Canny edge detector . Then, it invokes the HoughLines function which finds the lines, where each line is represented by the parameters . Then, we set an auxiliary function, PlotHoughLines , to plot the lines found by the Hough transform: The “classical” (non TensorFlow) version of the code is as follows: Computer Vision News Tool 9 Tool This code loads the Sudoku images (dave.png), extracts the lines in the image (getHoughlines) and plots them on the figure. The output of this process is in the figure at the right. “ It’s more than a deep learning package, it’s an architecture ”