: The filter "slides" to the right by a set distance (usually 1 pixel, known as a stride of 1) and repeats the calculation for the next set of inputs.
: The filter is placed over the first three elements of the input space. Each value in the filter is multiplied by the corresponding input value, and the results are summed to create a single output pixel . Example : If the input is [2, 4, 6] , the result is S W A T 1x3
: Specialized filters can detect sharp changes in intensity, effectively "outlining" shapes within data. : The filter "slides" to the right by
: In 2D images, a 1x3 filter specifically targets horizontal patterns, which can be useful for identifying textures or long horizontal lines. Real-Time Vision for Robot Swat-Juggling Example : If the input is [2, 4,
: The final sequence of these output values forms a feature map , which represents the detected presence and strength of the target pattern across the entire input. Practical Applications
In the context of , a 1x3 filter (or kernel) is a tool used for feature detection in a 1D input space or as a specific horizontal component in 2D image processing. How a 1x3 Filter Generates a Feature
: You define a filter with specific weights, such as [2, -1, 0.1] .