AttributeDistance

Attenuation function based distance from attributes

• file: distance/attribute.cpp
• inherits: UntrainableDistance
• author(s): Scott Klum
• properties: None

---

BayesianQuantizationDistance

Bayesian quantization Distance

• file: distance/bayesianquantization.cpp
• inherits: Distance
• author(s): Josh Klontz
• properties: None

---

ByteL1Distance

Fast 8-bit L1 distance

• file: distance/byteL1.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

CrossValidateDistance

Cross validate a Distance metric.

• file: distance/crossvalidate.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

DefaultDistance

DistDistance wrapper.

• file: distance/default.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

DistDistance

Standard Distance metrics

• file: distance/dist.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

EMDDistance

Computes Earth Mover's Distance https://www.cs.duke.edu/~tomasi/papers/rubner/rubnerTr98.pdf

• file: distance/emd.cpp
• inherits: UntrainableDistance
• author(s): Scott Klum
• properties: None

---

FilterDistance

Checks target metadata against filters.

• file: distance/filter.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

FuseDistance

Fuses similarity scores across multiple matrices of compared Template

• file: distance/fuse.cpp
• inherits: ListDistance
• author(s): Scott Klum

• properties:

  Property	Type	Description
  Operation	enum	Possible values are: Mean sum min max .

---

HalfByteL1Distance

Fast 4-bit L1 distance

• file: distance/halfbyteL1.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

HeatMapDistance

1v1 heat map comparison

• file: distance/heatmap.cpp
• inherits: Distance
• author(s): Scott Klum
• properties: None

---

IdenticalDistance

Returns true if the Template are identical, false otherwise.

• file: distance/identical.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

KeyPointMatcherDistance

Wraps OpenCV Key Point Matcher

• file: distance/keypointmatcher.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• see:
• properties: None

---

L1Distance

L1 distance computed using eigen.

• file: distance/L1.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

L2Distance

L2 distance computed using eigen.

• file: distance/L2.cpp
• inherits: UntrainableDistance
• author(s): Josh Klontz
• properties: None

---

MatchProbabilityDistance

Match Probability

• file: distance/matchprobability.cpp
• inherits: Distance
• author(s): Josh Klontz
• properties: None

---

MetadataDistance

Checks target metadata against query metadata.

• file: distance/metadata.cpp
• inherits: UntrainableDistance
• author(s): Scott Klum
• properties: None

---

NegativeLogPlusOneDistance

Returns -log(distance(a,b)+1)

• file: distance/neglogplusone.cpp
• inherits: Distance
• author(s): Josh Klontz
• properties: None

---

OnlineDistance

Online Distance metric to attenuate match scores across multiple frames

• file: distance/online.cpp
• inherits: UntrainableDistance
• author(s): Brendan klare
• properties: None

---

PermuteDistance

Compares all permutations of matrices from one template to the other, and fuses the scores via the operation specified.

• file: distance/permute.cpp
• inherits: Distance
• author(s): Scott Klum
• properties: None

---

PipeDistance

Distances in series.

The Template are compared using each Distance in order. If the result of the comparison with any given distance is -FLOAT_MAX then this result is returned early. Otherwise the returned result is the value of comparing the Template using the last Distance.

• file: distance/pipe.cpp
• inherits: ListDistance
• author(s): Josh Klontz
• properties: None

---

RejectDistance

Sets Distance to -FLOAT_MAX if a target Template has/doesn't have a key.

• file: distance/reject.cpp
• inherits: UntrainableDistance
• author(s): Scott Klum
• properties: None

---

SVMDistance

SVM Regression on Template absolute differences.

• file: distance/svm.cpp
• inherits: Distance
• author(s): Josh Klontz
• properties: None

---

SumDistance

Sum match scores across multiple Distance

• file: distance/sum.cpp
• inherits: ListDistance
• author(s): Scott Klum
• properties: None

---

ThresholdDistance

Applys a hard threshold to similarity values.

• file: distance/threshold.cpp
• inherits: Distance
• author(s): Scott Klum
• properties: None

---

TurkDistance

Unmaps Turk HITs to be compared against query mats

• file: distance/turk.cpp
• inherits: UntrainableDistance
• author(s): Scott Klum
• properties: None

---

UnitDistance

Linear normalizes of a Distance so the mean impostor score is 0 and the mean genuine score is 1.

• file: distance/unit.cpp
• inherits: Distance
• author(s): Josh Klontz
• properties: None

---

ZScoreDistance

Performs zscore normalization on distances at test time by learning mean and standard deviation parameters during training.

• file: distance/zscore.cpp
• inherits: Distance
• author(s): Scott Klum
• properties: None

---