We address the problem of uncertainty quantification in the domain of face attribute classification, using Evidential Deep Learning (EDL) framework. The proposed EDL approach leverages the strength of Convolution Neural Networks (CNN), with the objective of representing the uncertainty in the output predictions. Predominantly, the softmax/sigmoid activation functions are applied to map the output logits of the CNN to target class probabilities in multi-class classification problems. By replacing the standard softmax/sigmoid output of a CNN with the parameters of the evidential distribution, EDL learns to represent the uncertainty in its predictions. The proposed approach is evaluated on CelebA and LFWA datasets. The quantitative and qualitative analysis demonstrate the suitability and strength of EDL to estimate the uncertainty in the output predictions without hindering the accuracy of CNN-based models.
We propose a new model for learning to rank two images with respect to their relative strength of expression for a given attribute. We address this problem – called relative attribute learning — using a vision transformer backbone. The embedded representations of the two images to be compared are extracted and used for comparison with a ranking head, in an end-to-end fashion. The results demonstrate the strength of vision transformers and their suitability for relative attributes classification. Our proposed approach outperforms the state-of-the-art by a large margin, achieving 90.40% and 98.14% mean accuracy over the attributes of LFW-10 and Pubfig datasets.
We present a content-based image retrieval system for plant identification which is intended for providing users with a
simple method to locate information about their house plants. A plant image consists of a collection of overlapping leaves
and possibly flowers, which makes the problem challenging. We studied the suitability of various well-known color, texture
and shape features for this problem, as well as introducing some new ones. The features are extracted from the general
plant region that is segmented from the background using the max-flow min-cut technique. Results on a database of 132
different plant images show promise (in about 72% of the queries, the correct plant image is retrieved among the top-15
results).
User privacy and template security are major concerns in the use of biometric systems. These are serious concerns
based on the fact that once compromised, biometric traits can not be canceled or reissued. The Fuzzy Vault
scheme has emerged as a promising method to alleviate the template security problem. The scheme is based on
binding the biometric template with a secret key and scrambling it with a large amount of redundant data, such
that it is computationally infeasible to extract the secret key without possession of the biometric trait. It was
recently claimed that the scheme is susceptible to correlation based attacks which assume the availability of two
fuzzy vaults created using the same biometric data (e.g. two impressions of the same fingerprint) and suggests
that correlating them would reveal the biometric data hidden inside.
In this work, we implemented the fuzzy vault scheme using fingerprints and performed correlation attacks
against a database of 400 fuzzy vaults (200 matching pairs). Given two matching vaults, we could successfully
unlock 59% of them within a short time. Furthermore, it was possible to link an unknown vault to a short list
containing its matching pair, for 41% of all vaults. These results prove the claim that the fuzzy vault scheme
without additional security measures is indeed vulnerable to correlation attacks.
As biometrics gains popularity, there is an increasing concern about privacy and misuse of biometric data
held in central repositories. Furthermore, biometric verification systems face challenges arising from noise and
intra-class variations. To tackle both problems, a multimodal biometric verification system combining fingerprint
and voice modalities is proposed. The system combines the two modalities at the template level, using multibiometric
templates. The fusion of fingerprint and voice data successfully diminishes privacy concerns by hiding
the minutiae points from the fingerprint, among the artificial points generated by the features obtained from
the spoken utterance of the speaker. Equal error rates are observed to be under 2% for the system where 600
utterances from 30 people have been processed and fused with a database of 400 fingerprints from 200 individuals.
Accuracy is increased compared to the previous results for voice verification over the same speaker database.
The discriminative capability of a biometric is based on its
individuality/uniqueness and is an important factor in choosing a
biometric for a large-scale deployment. Individuality studies have
been carried out rigorously for only certain biometrics, in particular
fingerprint and iris, while work on establishing handwriting and signature individuality has been mainly on feature level.
In this study, we present a preliminary individuality model for online
signatures using the Fourier domain representation of the signature.
Using the normalized Fourier coefficients as global features describing the signature, we derive a formula for the probability of coincidentally matching a given signature. Estimating model parameters from a large database and making certain simplifying assumptions, the probability of two arbitrary signatures to match in 13 of the coefficients is calculated as 4.7x10-4. When compared with the results of a verification algorithm that parallels the theoretical model, the results show that the theoretical model fits the random forgery test results fairly well. While online signatures are sometimes dismissed as not very secure, our results show that the probability of successfully guessing an online signature is very low. Combined with the fact that signature is a behavioral biometric with adjustable complexity, these results support the use of online signatures for biometric authentication.
We describe a system for recognizing online, handwritten mathematical expressions. The system is designed with a user-interface
for writing scientific articles, supporting the recognition of basic mathematical expressions as well as integrals,
summations, matrices etc. A feed-forward neural network recognizes symbols which are assumed to be single-stroke and
a recursive algorithm parses the expression by combining neural network output and the structure of the expression.
Preliminary results show that writer-dependent recognition rates are very high (99.8%) while writer-independent symbol
recognition rates are lower (75%). The interface associated with the proposed system integrates the built-in recognition
capabilities of the Microsoft's Tablet PC API for recognizing textual input and supports conversion of hand-drawn
figures into PNG format. This enables the user to enter text, mathematics and draw figures in a single interface. After
recognition, all output is combined into one LATEX code and compiled into a PDF file.
In biometric based authentication, biometric traits of a person are matched against his/her stored biometric profile and access is granted if there is sufficient match. However, there are other access scenarios, which require participation of multiple previously registered users for a successful authentication or to get an access
grant for a certain entity. For instance, there are cryptographic constructs generally known as secret sharing schemes, where a secret is split into shares and distributed amongst participants in such a way that it is reconstructed/revealed only when the necessary number of share holders come together. The revealed secret can then
be used for encryption or authentication (if the revealed key is verified against the previously registered value). In this work we propose a method for the biometric based secret sharing. Instead of splitting a secret amongst participants, as is done in cryptography, a single biometric construct is created using the biometric traits of the participants. During authentication, a valid cryptographic key is released out of the construct when the required number of genuine participants present their biometric traits.
Despite recent developments in Tablet PC technology, there has not been any applications for recognizing handwritings in Turkish. In this paper, we present an online handwritten text recognition system for Turkish, developed using the Tablet PC interface. However, even though the system is developed for Turkish, the addressed issues are common to online handwriting recognition systems in general. Several dynamic features are extracted from the handwriting data for each recorded point and Hidden Markov Models (HMM) are used to train letter and word models. We experimented with using various features and HMM model topologies, and report on the effects of these experiments. We started with first and second derivatives of the x and y coordinates and relative change in the pen pressure as initial features. We found that using two more additional features, that is, number of neighboring points and relative heights of each point with respect to the base-line improve the recognition rate. In addition, extracting features within strokes and using a skipping state topology improve the system performance as well. The improved system performance is 94% in recognizing handwritten words from a 1000-word lexicon.
We describe a system for recognizing unconstrained Turkish handwritten text. Turkish has agglutinative morphology and theoretically an infinite number of words that can be generated by adding more suffixes to the word. This makes lexicon-based recognition approaches, where the most likely word is selected among all the alternatives in a lexicon, unsuitable for Turkish. We describe our approach to the problem using a Turkish prefix recognizer. First results of the system demonstrates the promise of this approach, with top-10 word recognition rate of about 40% for a small test data of mixed handprint and cursive writing. The lexicon-based approach with a 17,000 word-lexicon (with test words added) achieves 56% top-10 word recognition rate.
Recognition of general unconstrained cursive handwriting remains largely unsolved. We present a system for recognizing off-line cursive English text guided in part by global characteristics of the handwriting. A new method for finding the letter boundaries based on minimizing a heuristic cost function is introduced. The function is evaluated at each point along the baseline of the word to find the best possible segmentation points. The algorithm tries to find all the actual letter boundaries and as few additional ones as possible. After a normalization step that removes much of the style variation, the normalized segments are classified by a one hidden layer feedforward neural network. The word recognition algorithms find the segmentation points that are likely to be extraneous and generates all possible final segmentations of the word by either keeping or removing them. Interpreting the output of the neural network as posterior probabilities of letters, it then finds the word that maximizes the probability of having produced the image, over a set of 30,000 words and over all the possible final segmentations. We compared two hypotheses for finding the likelihood of words that are in the lexicon and found that using a Hidden Markov Model of English is significantly less successful than assuming independence among the letters of a word. In our initial test involving multiple writers, 68% of the words were in the top three choices.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.