Revision history [back]

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer

Hi, I'm using an algorithm in Knime wich do that :

(taken from source code of the node implemented in ImageLib)

• This class implements a very simple peak-picker, with optional ellipsoidal peak suppression. * Peaks are found by taking the sign of the difference operator in each dimension, differentiating * between negative and non-negative differences, then finding transitions from non-negative to * negative. This is accomplished in a random-access manner, in other words, with one * LocalizableCursor irrespective of how it traverses the {@link Image}, and a * LocalizableByDimCursor that is set to its 2^n-connected neighbors (where n is dimensionality).

i'can dound a way to implement that in opencv or better,, not reinventing the wheel THank you for your answer