# how last edge image can be achieved from Law masks

as we know thechnique of law edge detection has 25 2d masks that is obtained form 5 different 1d filters. we can use each of those masks on image with convolution but how can we obtain final result? is that achieved from from some kind of fusion or it have to be chosen?

thank you

• I'm not sure what you're talking about. Could you give a reference to the technique you're referring to? – Marcus Müller Jul 28 '17 at 20:47
• @marcus en.wikipedia.org/wiki/Image_texture searh for Laws Texture Energy Measures – virtouso Jul 28 '17 at 22:31
• OK, so it's "Laws Texture Energy Measures". The letters do make a difference. Also, I think the article is pretty well-explained and says exactly what you do – can you please be clearer by editing your question about what exactly you need help with? – Marcus Müller Jul 28 '17 at 22:36
• The source text  is pretty explicit: "a set of nine 5x5 convolution masks is used to compute the energy of texture which is then represented by a nine element vector for each pixel." it's that 9 element vector that describes the texture. Done - no step after. Section 7.3.4. There's a good reason that Wikipedia encourages people to cite their sources, and that is that interested and skeptic readers can read the original text. – Marcus Müller Jul 29 '17 at 7:54
• Is your question how to convolve an image with a mask? I really don't understand what your problem is. No, the result is not an image. If your original image was of size $n\times m$, the resulting structure is of shape $n\times m \times 9$. – Marcus Müller Jul 29 '17 at 15:20

## 1 Answer

If I understand correctly, the question is, given many images which are result of different Edge Filter applied on the same image, how to actually mark edges.

Well, you basically created 25 tests for each pixel to decide whether or not it is an edge.

You could apply many approaches to decide:

1. Majority Votes - If more than half of the voters decided it is an edge, nark it as an edge.
2. Threshold - If more than $x$ voters vote for edge the pixel will be declared as an edge.
3. Spatial Model - Instead of per pixel decision, look around it and other voters.
4. Weight of Votes - Don't mark votes as "Yes / No" but give it a scalar. If the sum of all scalars above a threshold value, declare an edge.

As you can see the options are endless.