Knowledge extraction can be done either through content-based or picture-based retrieval. With today’s Increasing popularity of image-based content, most searches are done through image-based queries, which will bring more accurate and relevant material based on the input. Picture-type queries differ from text-based queries in that the user requires domain knowledge. Before creating the image data sets, input or stored images must be trained. The user must make a designated similarity measure for the input set. It will help the researchers find relevant content more quickly and accurately. It will save searching and processing time. This paper employs several techniques, including using a colour histogram for the image, which aids in quantizing the image and calculating the colour of each pixel, which is saved for future processing. Using this histogram, the user creates different colour bins, which helps produce the desired outputs. The results of the trial validate this method. These techniques can be particularly beneficial in the context of Smart Home applications.