Edge Extraction: Gestures Alphabet of Colombian Signs.

The database presented consists of a set of images of the human hand making signs (20) at various angles, corresponding to the Colombian alphabet of signs established by the National Institute for the Deaf (INSOR). These signs are characterized by being static, that is, they do not require movement to be performed.

Signal processing is carried out by normalizing the data using the HSV (Hue - Saturation - Brightness) color space, using the "rgb2hsv" command in Matlab. In addition, noise removal is performed in each image using a Gaussian filter, which is a type of convolution used mainly to smooth images and remove noise, including black and white dots known as "salt and pepper". This process aims to obtain the necessary characteristics to carry out a thresholding step and a final edge extraction step.

This dataset is available for use and analysis, with the aim of contributing to the advancement in the field of gesture recognition and classification. It has been created with the purpose of being used in multiple artificial intelligence applications for pattern recognition and classification. For this reason, it has been divided into test and training datasets. This database is divided into two sub-folders (Training - Testing), each of which has folders for each gesture that each letter represents.