Ophthalmic visual field (static perimetry)

Read Ophthalmic Visual Field static-perimetry DICOM (Supplement 146, SOP Class 1.2.840.10008.5.1.4.1.1.80.1), flatten it to pandas or JSON, and check standard conformance. The semantic parsing runs in the native engine (the same reader behind pydcm.content); pydcm.opv is a thin pandas/JSON layer on top.

Read a single study

import pydcm

vf = pydcm.opv.read_dicom("vf.dcm")

vf.pointwise_to_pandas()       # one row per stimulus location
vf.to_pandas()                 # one row of study-level fields
vf.pointwise_to_nested_json()  # study identifiers + nested per-point records
vf.test_points                 # the raw per-point list

pointwise_to_pandas() returns a row per test point — x_coordinate, y_coordinate, sensitivity_value, stimulus_results, the age-corrected and generalized-defect deviations — each tagged with the study identifiers and a point_index. to_pandas() flattens the study-level groups (test_parameters, reliability, global_results) into one row.

Batch a directory

opvset, errors = pydcm.opv.read_dicom_directory("study_dir/")   # *.dcm by default

len(opvset)                       # OPV files that parsed
errors                            # [(path, message), ...] for non-OPV / unreadable
opvset.pointwise_to_pandas()      # every file's points, concatenated
opvset.to_pandas()                # one study-level row per file
opvset.check_dicom_compliance()   # {path: [findings, ...]}

Check conformance

vf.check_dicom_compliance()       # IOD / module findings for this file

This reuses the native IOD conformance judge (pydcm.iod_validate), which enforces every mandatory module's Type-1 / Type-2 attribute presence for the OPV SOP Class, descending into present sequences. It is stricter than a flat tag checklist — per-SOP-Class and nested-sequence aware — so an empty list means the file is conformant at the IOD level.

import pydcm
pydcm.iod_validate("vf.dcm")      # [{severity, tag, module, message}, ...]

The semantic content directly

pydcm.opv.read_visual_field (or pydcm.content) returns the raw nested content — useful when you want the structure without pandas:

import pydcm

c = pydcm.content("vf.dcm")
c["type"]                     # "ophthalmic_visual_field"
c["laterality"]               # "OD" / "OS"
c["test_parameters"]          # extent, shape, stimulus luminance/area/time, ...
c["reliability"]              # fixation losses, false +/-, catch trials, ...
c["global_results"]           # mean sensitivity, MD/PSD-family deviations,
                              # short-term fluctuation, blind spot, ...
c["test_points"]              # per-stimulus records

What is extracted

Every scalar and coded field of the perimetry measurements is surfaced — test parameters, reliability, the global-result deviations, the blind spot, and per-point measurements with their deviation probabilities. The deep normative / algorithm reference sequences (e.g. TestPointNormalsSequence, AgeCorrectedSensitivityDeviationAlgorithmSequence) are reported by presence rather than expanded.