Section outline

  • Image Kernels

    🎙️Marcelo Gutiérrez; Claudio Acuña


    An image kernel is a tiny matrix that is used to create effects similar to those seen in Photoshop or Gimp, such as blurring, sharpening, outlining, and embossing. They are also used in machine learning for 'feature extraction,' which is a method for detecting the most relevant parts of a picture. In this context, the process is known more broadly as convolutional neural networks (ConvNet).


    Sobel filter

    Edges are abrupt discontinuities in an image that can be caused by surface color, depth, illumination, or other factors. In this sense, we identify the edges in the image below wherever there is a sudden change in intensity.