| Cmp Name | Parameters | Description |
| ccc | emdata mask: image mask int negative: If set, returns -1 * ccc product. Set by default so smaller is better |
Cross-correlation coefficient (default -1 * ccc) |
| ccc.tomo | emdata ccf: The ccf image, can be provided if it already exists to avoid recalculating it bool norm: Whether the cross correlation image should be normalized (should be for normalized images). Default is true. emdata normalize: Return the negative value (which is EMAN2 convention), Defalut is true(1) int searchx: The maximum range of the peak location in the x direction. Default is sizex/4 int searchy: The maximum range of the peak location in the y direction. Default is sizey/4 int searchz: The maximum range of the peak location in the z direction. Default is sizez/4 |
Ccc with consideration given for the missing wedge |
| ccc.tomo.thresh | int negative: If set, returns -1 * ccc product. Set by default so smaller is better float sigmaimg: Sigma coefficient for thresholding values included in the dot product. default = 0.5 float sigmawith: Sigma coefficient for thresholding values included in the dot product in the 'with' image. Default = 0.5 |
Ccc of two FFTs with missing wedge taken into account explicitly. Also stores overlap as 'fft_overlap' in first input image. Individual voxels are include only if they exceed the threshold value in both images. fft_overlap expresses the fractional number of voxels included in the comparison. Actual returned correlations already account for this effect, but the overlap can be used to estimate the relative uncertainty in the resulting correlation coefficient. |
| dot | emdata mask: image mask int negative: If set, returns -1 * dot product. Set by default so smaller is better int normalize: If set, returns normalized dot product (cosine of the angle) -1.0 - 1.0. |
Dot product (default -1 * dot product) |
| frc | int ampweight: If set, the amplitude of 'this' will be used to weight the result (default=0) float maxres: Highest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=10 float minres: Lowest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=500 int nweight: Downweight similarity based on number of particles in reference (default=0) float pmax: The maximum resolution in pixels. float pmin: The minimum resolution in pixels. int snrweight: If set, the SNR of 'this' will be used to weight the result. If 'this' lacks CTF info, it will check 'with'. (default=0) int sweight: If set, weight the (1-D) average by the number of pixels in each ring (default=1) int zeromask: Treat regions in either image that are zero as a mask (default=0) |
Computes the mean Fourier Ring Correlation between the image and reference (with optional weighting factors). |
| fsc.tomo | float apix: The angstroms per pixel to use. Default = apix_x(1.0 if not present) float maxres: The maximum resolution to accept (1/A) Default=0.0 float minres: The minimum resolution to accept (1/A) Default is inf emdata normalize: Return the negative value (which is EMAN2 convention), Defalut is true(1) float sigmas: The number of times the standard deviation of Fourier amplitudes to accept |
EXPERIMENTAL - Fsc with consideration given for the missing wedge. Not ready for routine use. |
| fsc.tomo.auto | float maxres: The maximum resolution to accept (1/A) Default is axial Nyquist. overwrites pmax float minres: The minimum resolution to accept (1/A) Default is 3 pixels. overwrites pmin int pmax: The maximum resolution in pixels. int pmin: The minimum resolution in pixels. bool retcurve: Return fsc curve in image header floatarray sigmaimg: Res dependent amplitude threshold. If provided this must contain a floating point array with as many elements as the Fourier radius of the image. float sigmaimgval: Sigma coefficient for thresholding values included in the dot product. default = 0.5 floatarray sigmawith: Resolution dependent coefficient for thresholding values included in the dot product in the 'with' image. Default = 0.1 and is normally fine float sigmawithval: Sigma coefficient for thresholding values included in the dot product in the 'with' image. Default = 0.5 |
FSC of two FFTs with missing wedge taken into account via a dynamic threshold. Also stores overlap as 'fft_overlap' in first input image on return. This comparator can take the missing wedge into account automatically without any precomputation of wedge orientation. If sigmaimg and/or sigmawith are provided, they must contain floating point arrays with at least as many elements as the Fourier radius in pixels, and these curves represent the thresholds for each image for including the values in the FSC, or considering them to be missing values. If not provided, these values will be computed automatically. These options are primarily provided for speed. |
| lod | emdata mask: image mask int negative: If set (which is the default), returns Lod. (The smaller the better) int normalize: If set, normalizes maps prior to computing the difference. Default=0 (no normalization) |
L^1 normalized difference (positive by default) |
| optsub | bool ctfweight: Weight frequencies using CTF emdata mask: Real space mask. Only computes the residual power under the mask. Significant speed penalty if specified. Default=None. float maxres: Highest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=10 float minres: Lowest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=200 bool zeromask: Treat zero regions in 'with' as a mask |
Residual power left in the image after optimally subtracting the reference with math.sub.optimal. Smaller values indicate a closer match. |
| optvariance | int debug: Performs various debugging actions if set. int invert: If set, 'with' is rescaled rather than 'this'. 'this' should still be the noisier image. (default=0) int keepzero: If set, zero pixels will not be adjusted in the linear density optimization. (default=1) int matchamp: Takes per-pixel Fourier amplitudes from self and imposes them on the target, but leaves the phases alone. (default=0) int matchfilt: If set, with will be filtered so its radial power spectrum matches 'this' before density optimization of this. (default=1) int radweight: Upweight variances closer to the edge of the image. (default=0) |
Real-space variance after density optimization, self should be noisy and target less noisy. Linear transform applied to density to minimize variance. |
| phase | int ampweight: If set, the amplitude of 'with' will be used as a weight in the averaging'. (default=0) float maxres: Highest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=10 float minres: Lowest resolution to use in comparison (soft cutoff). Requires accurate A/pix in image. <0 disables. Default=500 int snrfn: If nonzero, an empirical function will be used as a radial weight rather than the true SNR. (1 - exp decay)'. (default=0) int snrweight: If set, the SNR of 'this' will be used to weight the result. If 'this' lacks CTF info, it will check 'with'. (default=0) int zeromask: Treat regions in either image that are zero as a mask |
Mean phase difference |
| quadmindot | int negative: If set, returns -1 * dot product. Default = true (smaller is better) int normalize: If set, returns normalized dot product -1.0 - 1.0. |
Calculates dot product for each quadrant and returns worst value (default -1 * dot product) |
| sqeuclidean | emdata mask: image mask int normto: If set, 'with' is normalized to 'this' before computing the distance int zeromask: If set, zero pixels in either image will be excluded from the statistics |
Squared Euclidean distance (sum(a - b)^2)/n. |
| vertical | Align the 2d image vertically. Reference image is ignored |