vault backup: 2024-10-02 23:25:27

Biometric Systems/final notes/Untitled.md

@@ -0,0 +1,99 @@
+#### Problems of biometric systems:
+- wide intra-class variations
+	- maybe different facial expression, different light, different view point...
+- very small inter-class variations
+	- two different person very similar (i.e. twins)
+
+- possible spoofing attacks, in different moments
+	![[Pasted image 20241002181936.png]]
+
+- [non universality](LEZIONE2_Indici_di_prestazione.pdf#page=6&selection=0,10,0,26&color=yellow|LEZIONE2_Indici_di_prestazione, p.6)
+	- e.g. people with no voice, people with cataract, people with poor fingerprints...
+
+Most difficult traits to exploit:
+- retina fundus
+- behavioral traits (i.e. way of walking)
+- handwriting
+
+### [What to compare?](LEZIONE2_Indici_di_prestazione.pdf#page=8&selection=0,10,0,26&color=yellow|LEZIONE2_Indici_di_prestazione, p.8)
+- **Sample
+	- raw captured data
+- **Hand-crafted features**
+	- manually engineered by the data scientist and extracted from samples
+	- can also be substituted with **embeddings**: features automatically extracted by deep architectures
+- **Template**
+	- collection of features extracted from the row data, examples:
+		- a histogram representing the frequencies of relevant values in the image (e.g. greylevel values)
+		- a vector of values each representing a relevant measure (e.g. Bertillon measures)
+		- time series of acceleration values (one per axis)
+		- a set of triplets as for relevant fingerprint points representing the coordinates of the points and the direction of the tangent to the ridge in that point.
+
+
+> [!PDF|red] [[LEZIONE2_Indici_di_prestazione.pdf#page=8&selection=11,1,14,16&color=red|LEZIONE2_Indici_di_prestazione, p.8]]
+> > Hand-crafted features
+> 
+> not the template of the entire biometric system.
+
+
+### Comparing templates
+- Euclidian distance
+- Cosine similarity
+	- cosine of the angle between two vectors
+- Pearson correlation
+- Bhattacharyya distance
+
+
+> [!PDF|yellow] [[LEZIONE2_Indici_di_prestazione.pdf#page=9&selection=8,0,10,31&color=yellow|LEZIONE2_Indici_di_prestazione, p.9]]
+> > or cosine similarity may provide either a distance measure or a similarity measure
+> 
+> shows "more stuff" than Euclidian distance, such as orientation ecc.. Shows how templates are similar to eachother. While distance shows how templates are... distant!
+
+> [!PDF|yellow] [[LEZIONE2_Indici_di_prestazione.pdf#page=10&selection=3,1,4,21&color=yellow|LEZIONE2_Indici_di_prestazione, p.10]]
+> > (Pearson) Correlation
+> 
+> how signals are similar to eachother. Often used to compare fingerprints, by computing the correlation between two fingerprints.
+
+Histograms needs other ways to be compared.
+
+The same happens with time series: speed for example may speed the final outcome of the time series, even if the patterns are the same.
+
+So what do we do?
+sometimes we use correlation, but Dynamic time Warping is the most used.
+
+![[Pasted image 20241002135922.png]]
+
+each point is paired with the most convenient one. It's not necessaty that points corresponds to the same instant in time.
+
+if using deep learning we should use the architecture to extract the embeddings (for both gallery and probe templates).
+
+//
+after normalization in range [0, 1] we will have that distance = 1 - similarity.
+
+#### Possible errors: verification
+- Genuine Match (GM, GA): the claimed identity is true and subject is accepted
+- False Rejection (FR, FNM, type I error): claimed identity is true but the subjet is rejected
+- Genuine Reject (GR, GNM): an impostor is rejected
+- False Acceptance (FA, FM, type II error): an impostor is accepted :/
+
+It's important to define a good threshold.
+If too high we will get a lot of false acceptance. If too low we will get a lot of false rejection!
+
+When computing rates:
+- False Rejection Rate (FRR) is the number of FR divided by ONLY the number of GM+FR.
+	- in fact, GM + FR have the same denominator and sum up to 1.
+- False Acceptance Rate is the number of FA divided by FA + GR
+
+Equal Error Rate is the value at a specific threshold, where FAR and FRR are the same value.
+
+two synthetic metrics could be ERR and area below ROC curve.
+
+(we might have more templates for the same person to address inter-class variation.
+Of course templates should be different, not computed i.e. by frames of the same video, as some of them could be blurred and close frames are exactly the same!)
+
+> [!PDF|yellow] [[LEZIONE2_Indici_di_prestazione.pdf#page=20&selection=119,0,119,4&color=yellow|LEZIONE2_Indici_di_prestazione, p.20]]
+> > When
+> 
+> in false acceptance we can have two possible scenarios
+> - pj does not belong to the gallery (most trivial)
+> - pj belongs to an enrolled subject but the probe claimed another identity, not the real one.
+