IncrementalCorrelationFilterThinWrapper¶
-
class
menpofit.clm.
IncrementalCorrelationFilterThinWrapper
(cf_callable=<function mccf>, icf_callable=<function imccf>)[source]¶ Bases:
object
Wrapper class for defining an Incremental Correlation Filter.
Parameters: - cf_callable (callable, optional) –
The correlation filter function. Possible options are:
Class Method mccf
Multi-Channel Correlation Filter mosse
Minimum Output Sum of Squared Errors Filter - icf_callable (callable, optional) –
The incremental correlation filter function. Possible options are:
Class Method imccf
Incremental Multi-Channel Correlation Filter imosse
Incremental Minimum Output Sum of Squared Errors Filter
-
increment
(A, B, n_x, Z, t)[source]¶ Method that trains the correlation filter.
Parameters: - A (
(N,)
ndarray) – The current auto-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- B (
(N, N)
ndarray) – The current cross-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- n_x (int) – The current number of images.
- Z (list or
(n_images, n_channels, patch_h, patch_w)
ndarray) – The training images (patches). If list, then it consists of n_images(n_channels, patch_h, patch_w)
ndarray members. - t (
(1, response_h, response_w)
ndarray) – The desired response.
Returns: - correlation_filter (
(n_channels, response_h, response_w)
ndarray) – The learned correlation filter. - auto_correlation (
(N,)
ndarray) – The auto-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- cross_correlation (
(N, N)
ndarray) – The cross-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- A (
-
train
(X, t)[source]¶ Method that trains the correlation filter.
Parameters: - X (list or
(n_images, n_channels, patch_h, patch_w)
ndarray) – The training images (patches). If list, then it consists of n_images(n_channels, patch_h, patch_w)
ndarray members. - t (
(1, response_h, response_w)
ndarray) – The desired response.
Returns: - correlation_filter (
(n_channels, response_h, response_w)
ndarray) – The learned correlation filter. - auto_correlation (
(N,)
ndarray) – The auto-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- cross_correlation (
(N, N)
ndarray) – The cross-correlation array, whereN = (patch_h+response_h-1) * (patch_w+response_w-1) * n_channels
- X (list or
- cf_callable (callable, optional) –