I wrote this Code fragment
Which is a E (quarter) note in Treble Clef. Instead Mathematica simply classified it as a "musical note" Can we do better?
$\begingroup$
$\endgroup$
4
1 Answer
$\begingroup$
$\endgroup$
7
Here is my humble approach.
sheet = Import["https://rmurta.s3-sa-east-1.amazonaws.com/mathematica-playground/sheet.png"];
note = Import["https://rmurta.s3-sa-east-1.amazonaws.com/mathematica-playground/note.png"];
examples = Map[ImageResize[#, 100]&]/@ <|
"mi" -> (ImageCompose[sheet, note,{200,235} + #]&/@RandomReal[{-5,5},{4,2}]),
"fa" -> (ImageCompose[sheet, note,{200,280} + #]&/@RandomReal[{-5,5},{4,2}]),
"sol" ->(ImageCompose[sheet, note,{200,360} + #]&/@RandomReal[{-5,5},{4,2}]),
"la" ->(ImageCompose[sheet, note,{200,400} + #]&/@RandomReal[{-5,5},{4,2}]),
"si" ->(ImageCompose[sheet, note,{200,430} + #]&/@RandomReal[{-5,5},{4,2}]),
"do" ->(ImageCompose[sheet, note,{200,490} + #]&/@RandomReal[{-5,5},{4,2}])
|>;
c = Classify[examples];
Copying your example into Mathematica, and applying c into It:
-
$\begingroup$ Thanks but did not work for any other notes.. $\endgroup$ May 5, 2020 at 7:32
-
$\begingroup$ Can you please send me the images of list of notes you have that worked for you.. $\endgroup$ May 5, 2020 at 14:56
-
$\begingroup$ Just execute:
ImageCompose[sheet, note,{200,360}]$\endgroup$– MurtaMay 5, 2020 at 15:38 -
$\begingroup$ I thought as much. This is unrelated to Machine Learning. For example, it does not work for handwritten notes. I wanted a machine-learning algorithm that can identify notes the way a human would. $\endgroup$ May 6, 2020 at 8:30
-
$\begingroup$ I want to use commands like ImageIdentify that are machine learning related. $\endgroup$ May 6, 2020 at 8:31
ImageIdentify) can do is identify it as a musical note. I look at it in the context of a bass clef and see either a low G, G-sharp, or G-flat. $\endgroup$