Skip to content

Setting file description

Serdar Tasel edited this page Dec 3, 2024 · 1 revision

MitoSeg provides the flexibility to customize additional settings for fine-tuning the algorithm through a configuration file in YAML format. Below is a brief description of the customizable items:

General settings

NUM_CORES: The number of CPU cores utilized

THREAD_STACK_SIZE: Size of allocated stack in MB in CPU threads

Phase 1 Parameters

Visualization parameters

VISUAL_LE_SSIZE: Size of blocks in local energy map image used for visualization purposes

Preprocessing parameters

TFACTOR: Pixel size in nm for resampling. A pixel size of 2 nm appears to be sufficient to visualize membranes in detail.

AUTOLEVELSCUT: The histogram cut parameter of auto-contrast adjustment step. This should be lowered as small as possible since a high threshold may lead to data loss.

BSMOOTH_GRAY: Bilateral filtering intensity sigma parameter

Common curve fit parameters

FC_MAXCURV: Max. curvature parameter

FC_MAXITER: Max. number of iterations allowed

FC_STEP: Distance between consecutive sampled points in parabolic arcs

FC_XYRANGE: Neighborhood size determining the search space for tip points parameter of parabolic arcs

FC_XYSTEP: Step size in the search space for tip points parameter of parabolic arcs

FC_HRANGE: Neighborhood size determining the search space for height parameter of parabolic arcs

FC_HSTEP: Step size in the search space for height parameter of parabolic arcs

LE_BINS: Angle quantization parameter for local energy mapping process

Low frequency curve fit parameters

FCS_INIT_THRESH_LO: Min. ridge strength criterion in local energy map to extract large-scale (low frequency) curves

High frequency curve fit parameters

FCS_INIT_THRESH_HI: Min. ridge strength criterion in local energy map to extract small-scale (high frequency) curves

Low frequency curves coverage distance

CURV_HILO_COV_TH: Min. percentage threshold to detect overlapping small and large-scale curves

Detection parameters

DT_BOUNDARY_T: Min. threshold for average boundary strength of candidate mitochondria

DT_REGION_T: Min. threshold for average cristae strength of candidate mitochondria

DT_GAP_T: Min. threshold for boundary point strength to be considered as a gap

DT_GAPTOTAL_TD: Max. allowed total gap size (in nm) for candidate mitochondria

DT_GAPMAX_TD: Max. allowed size (in nm) of each boundary gap for candidate mitochondria

DT_GAPTOTALRATIO_T: Max. allowed gap ratio for candidate mitochondria

DT_GAPBORDERRATIO_T: Max. allowed gap ratio (on the image border) for candidate mitochondria

DT_GAPBORDERCOUNT_T: Max. allowed number of borders to eliminate candidate mitochondria appearing image border (1: Mitochondria boundary on at least one of top/bottom/left/right border, 2: Mitochondria boundary on at least two border, e.g. top-left)

DT_CURVG_T: Max. allowed global curvature calculated over the whole mitochondria boundary

DT_CURVL_T: Max. allowed local curvature on mitochondria boundary

DT_CURVL_SEG: The ratio of boundary segment used to calculate local curvature

DT_SIGN_SMOOTH: The signature smoothing parameter

DT_SIGN_RATIO: Min. ratio of boundary necessary to detect local extrema in the signature function

DT_SIGN_MAXNUMCRIT_T: Max. number of local extrema in the signature function allowed for candidate mitochondria

DT_MINMINOR_RATIO_T: Min. allowed ratio of the minimum cross-sectional thickness to the minor axis size

DT_REPORT: If true, the detection status will be reported in the terminal for each candidate mitochondrion

DT25D_MED_RANGE: The window size in z-axis to compute the median boundary strength

Phase 2 Parameters

Common snake parameters

SN_N: The number of points in the snake model

SN_GAUSSIAN: Gaussian smoothing sigma parameter for calculating curve energy

2.5D snake parameters

SN25D_W_TENSION: The 2D tension weight used in the snake model

SN25D_W_CURVATURE: The 2D curvature weight used in the snake model

SN25D_W_ZTENSION: The z-axis tension weight used in the snake model

SN25D_W_ZCURVATURE: The z-axis curvature weight used in the snake model

SN25D_W_ECURVE: The curve energy weight used in the snake model

SN25D_W_EINF_MIN: The min. inflation force weight used in the snake model

SN25D_W_EINF_MAX: The max. inflation force weight used in the snake model

SN25D_W_EINF_STEP: The step size of inflation force weight

SN25D_T: The z-thickness parameter of the snake model

SN25D_K: The snake update parameter

SN25D_INITR: The initial snake radius

SN25D_MAXITER: The maximum iterations allowed for snake fitting algorithm

SN25D_SHORTTERM_CONV_ITER: The number of iterations used to calculate short-term convergence

SN25D_SHORTTERM_CONV: The maximum movement for a snake vertex to consider as converged in short-term

SN25D_LONGTERM_CONV_ITER: The number of iterations used to calculate long-term convergence

SN25D_LONGTERM_CONV: The maximum movement for a snake vertex to consider as converged in long-term

SN25D_INF_CONV: The similarity threshold to ensure that only unique snakes are extracted

Phase 3 Parameters

POLY_MERGE: The threshold for the ratio of overlapping regions to entire region of snake outputs to start merging

POLY_VALIDITY: The threshold used to validate the snake output