EP0731961B1
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Money validator
Assignee: Mars Inc
•
Filing: 1994-11-30
•
Granted: 1998-05-20
•
Expires: 2014-11-30
Abstract
No abstract record available.
Inventors
- 👤 Gaston Baudat
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 18:20
CPC Classifications
G; G07; G07D; G07D7/00; G07D7/06; G07D7/12; G06; G06F; G06F18/00; G06F18/20; G06F18/24; G07D7/121
Claims Analysis
Total Claims: 17
Claim 1
Independent Claim
134 words
A method of classifying a test article as one of a
plurality of acceptable denominations of articles of currency,
comprising: deriving a plurality of measurement signals from a test
article; and applying an acceptance process to said measurement
signals, said acceptance process calculating first, second and
third outputs (382a-382f) indicating correspondence with
first, second and third denominations, characterised in that
for each said test item said acceptance process is performed
a plurality of times, differently adjusted to compensate for
the effects of measurement drift for at least said first,
second and third denominations at each performance, and in
that in each performance of said acceptance process, an output
corresponding to the denomination for which the acceptance
process is adjusted is calculated, and at least two other
outputs for two different denominations are also calculated.
Claim 2
Dependent Claim
19 words
A method according to claim 1 in which said acceptance
process is performed said plurality of times in succession.
Claim 3
Dependent Claim
45 words
A method according to claim 1 or claim 2 in which said
acceptance process is adjusted by processing said measurements
using, for each denomination, stored drift data (σ), to
produce corrected measurements for each denomination, and
applying a common acceptance process to said corrected
measurements.
Claim 4
Independent Claim
96 words
A method according to any preceding claim in which said
acceptance process employs stored reference data which has
previously been derived from training articles known to be
valid or invalid examples of said denominations and is
arranged to distinguish therebetween, in which level data
relating to the range or average level of measurement values
derived from said training articles is stored, for each of
said denominations, and in which measurements derived from
test articles are processed using said stored level data to
align them with each said stored range for each denomination
during said classification process.
Claim 5
Dependent Claim
21 words
A method according to claim 4, in which said level data
is stored as an average value for each said denomination.
Claim 6
Dependent Claim
36 words
A method according to claim 4 or claim 5 in which said
processing comprises forming, for each denomination, a
correction factor responsive to the ratio between said level
data and a running average of said measurements.
Claim 7
Independent Claim
42 words
A method according to any preceding claim characterised
by applying also an acceptance boundary test which limits the
acceptance volume for each denomination so as to exclude
forgeries (F1, F2 ...) not corresponding to training articles
from which said process was derived.
Claim 8
Dependent Claim
46 words
A method according to claim 7 in which the acceptance
boundary test comprises applying, for each acceptable denomination, acceptance
criteria which are relatively easily met so as to reject those
denominations for which said relaxed acceptance criteria are
not met prior to applying said acceptance process.
Claim 9
Independent Claim
38 words
A method according to any preceding claim in which the
acceptance process comprises forming, for each acceptable
denomination, a linear function of said measurement signals;
and applying a continuous non-linear function to the result
of said linear function.
Claim 10
Independent Claim
17 words
A method according to any preceding claim, in which the
acceptance process comprises a neural network process.
Claim 11
Dependent Claim
32 words
A method according to claim 10 in which the neural
network process comprises a feed-forward process, the
constants characterising the process having been iteratively
derived in a training phase by back propagation.
Claim 12
Dependent Claim
19 words
A method according to claim 10 or claim 11 in which the
neural network process comprises two network layers.
Claim 13
Dependent Claim
29 words
A method according to claim 12 in which the hidden layer
of said neural network comprises substantially fewer (e.g. an
order of magnitude less) nodes than the inputs thereto.
Claim 14
Independent Claim
16 words
A method according to any preceding claim
employing optical sensing means for deriving the measurement signals.
Claim 15
Dependent Claim
19 words
A method according to claim 14 in which the
optical sensing is performed at a plurality of different
wavelengths.
Claim 16
Dependent Claim
16 words
Use of a method according to any preceding claim
in a money validator, e.g. for banknotes.
Claim 17
Dependent Claim
149 words
A money validator for classifying a test article as one of a plurality of
acceptable denominations of articles of currency, comprising: measuring means for deriving a plurality of measurement signals from a
test article; and means for applying an acceptance process to said measurement signals,
comprising calculating means for calculating first, second and third outputs
(382a-382f) indicating correspondence with first, second and third
denominations, characterised in that means are provided for automatically
performing, for each said test item, said acceptance process a plurality of times,
differently adjusted by adjusting means to compensate for the effects of
measurement drift for at least said first, second and third denominations at each
performance, and in that in each performance of said acceptance process, the
calculating means calculates an output corresponding to the denomination for
which the acceptance process is adjusted, and also calculates at
least two other outputs for two different denominations.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.