Interface: FrameSummary
@kortexya/reasoninglayer / FrameSummary
Interface: FrameSummary
Section titled “Interface: FrameSummary”Defined in: src/types/vision.ts:61
Per-frame Ψ-term prediction summary.
Properties
Section titled “Properties”frameIndex
Section titled “frameIndex”frameIndex:
number
Defined in: src/types/vision.ts:66
Zero-based index in the temporal token grid. It is not necessarily an
index of the original video; see sourceFrameIndices.
numPredictedObjects
Section titled “numPredictedObjects”numPredictedObjects:
number
Defined in: src/types/vision.ts:68
Number of predicted objects for this frame.
numResiduatedFeatures
Section titled “numResiduatedFeatures”numResiduatedFeatures:
number
Defined in: src/types/vision.ts:70
Number of residuated features for this frame.
objectLabels
Section titled “objectLabels”objectLabels:
string[]
Defined in: src/types/vision.ts:72
Per-object predicted-sort labels, top-1 only.
osfVerdicts
Section titled “osfVerdicts”osfVerdicts:
OsfVerdict[]
Defined in: src/types/vision.ts:74
Per-object declared-obligation verdicts, aligned with objectLabels.
paddedInputIndices
Section titled “paddedInputIndices”paddedInputIndices:
number[]
Defined in: src/types/vision.ts:76
Indices of the model input frames of this grid step that are padding.
sourceFrameIndices
Section titled “sourceFrameIndices”sourceFrameIndices:
number[]
Defined in: src/types/vision.ts:78
Indices of the decoded source-video frames this grid step covers.
sourceTimestampsUs
Section titled “sourceTimestampsUs”sourceTimestampsUs: (
number|null)[]
Defined in: src/types/vision.ts:83
Source timestamps in microseconds, aligned with sourceFrameIndices;
null where the container gave none.