Filters








6,689 Hits in 3.8 sec

Understanding Unequal Gender Classification Accuracy from Face Images [article]

Vidya Muthukumar, Tejaswini Pedapati, Nalini Ratha, Prasanna Sattigeri, Chai-Wah Wu, Brian Kingsbury, Abhishek Kumar, Samuel Thomas, Aleksandra Mojsilovic, Kush R. Varshney
2018 arXiv   pre-print
Recent work shows unequal performance of commercial face classification services in the gender classification task across intersectional groups defined by skin type and gender.  ...  A second suspect, hair length, is also shown not to be the driver via experiments on face images cropped to exclude the hair.  ...  Introduction The problem of unequal accuracy rates across groups has recently been highlighted in gender classification from face images.  ... 
arXiv:1812.00099v1 fatcat:v2ofx5uzfjeaxfk5mppzi2gzzi

Understanding Fairness of Gender Classification Algorithms Across Gender-Race Groups [article]

Anoop Krishnan, Ali Almadan, Ajita Rattani
2020 arXiv   pre-print
Middle Eastern males and Latino females obtained higher accuracy rates most of the time. Training set imbalance further widens the gap in the unequal accuracy rates across all gender-race groups.  ...  Specifically, the majority of the studies raised the concern of higher error rates of the face-based gender classification system for darker-skinned people like African-American and for women.  ...  Muthukumar [9] analyzed the influence of the skin type for understanding the reasons for unequal gender classification accuracy on face images.  ... 
arXiv:2009.11491v1 fatcat:7x5fjnsn5bh3lfaywwrdjo3vqe

An Examination of Bias of Facial Analysis based BMI Prediction Models [article]

Hera Siddiqui, Ajita Rattani, Karl Ricanek, Twyla Hill
2022 arXiv   pre-print
Research on bias evaluation of face-based gender-, age-classification, and face recognition systems suggest that these technologies perform poorly for women, dark-skinned people, and older adults.  ...  Most of these studies used BMI annotated facial image datasets that mainly consisted of Caucasian subjects.  ...  Tab. 5 shows the average classification accuracy values obtained by the five models across the four gender-race groups. The overall accuracy ranged from 56.79% to 61.98%.  ... 
arXiv:2204.10262v1 fatcat:s7r3tewzgrgszetm4nyvmxrwb4

Single Unit Status in Deep Convolutional Neural Network Codes for Face Identification: Sparseness Redefined [article]

Connor J. Parde, Y. Ivette Colón, Matthew Q. Hill, Carlos D. Castillo, Prithviraj Dhar, Alice J. O'Toole
2020 arXiv   pre-print
Deep convolutional neural networks (DCNNs) trained for face identification develop representations that generalize over variable images, while retaining subject (e.g., gender) and image (e.g., viewpoint  ...  Gender classification declined gradually and viewpoint estimation fell steeply as dimensionality decreased.  ...  In total, each AUC was computed from 30,935,844 comparisons. Classification Gender Linear discriminant analysis (LDA) was used to classify face gender for each image in the dataset.  ... 
arXiv:2002.06274v2 fatcat:qzk2sjgskbhoxfkegvghnykchu

InclusiveFaceNet: Improving Face Attribute Detection with Race and Gender Diversity [article]

Hee Jung Ryu, Hartwig Adam, Margaret Mitchell
2018 arXiv   pre-print
The system, which we call InclusiveFaceNet, detects face attributes by transferring race and gender representations learned from a held-out dataset of public race and gender identities.  ...  We demonstrate an approach to face attribute detection that retains or improves attribute detection accuracy across gender and race subgroups by learning demographic information prior to learning the attribute  ...  Here, X S and X D are completely disjoint sets, respectively representing face images for face attribute detection and face images for race classification.  ... 
arXiv:1712.00193v3 fatcat:h5oq2baa4rbhliccezomis7cxi

Algorithmic Discrimination: Formulation and Exploration in Deep Learning-based Face Biometrics [article]

Ignacio Serna, Aythami Morales, Julian Fierrez, Manuel Cebrian, Nick Obradovich, Iyad Rahwan
2019 arXiv   pre-print
The experiments are conducted over the new DiveFace database composed of 24K identities from six different demographic groups.  ...  The main aim of this study is focused on a better understanding of the feature space generated by deep models, and the performance achieved over different demographic groups.  ...  Automatic classification algorithms based on these three categories show performances of up to 98% accuracy (Morales, Fierrez, and Vera-Rodriguez 2019).  ... 
arXiv:1912.01842v1 fatcat:dnb473xvj5fozg2jlyesatyaum

Learning Emotional-Blinded Face Representations [article]

Alejandro Peña and Julian Fierrez and Agata Lapedriza and Aythami Morales
2020 arXiv   pre-print
We show that it is possible to eliminate information related to emotion recognition tasks, while the performance of subject verification, gender recognition, and ethnicity classification are just slightly  ...  The results demonstrate that it is possible to reduce emotional information in the face representation while retaining competitive performance in other face-based artificial intelligence tasks.  ...  (k = 1: for Gender classification, k = 2: Ethnicity classification, k = 3: Emotion classification).  ... 
arXiv:2009.08704v1 fatcat:l3ro5hluujhlbi62q3s2r2ykhu

Closing the gap between single-unit and neural population codes: Insights from deep learning in face recognition

Connor J. Parde, Y. Ivette Colón, Matthew Q. Hill, Carlos D. Castillo, Prithviraj Dhar, Alice J. O'Toole
2021 Journal of Vision  
Gender and viewpoint classification required large-scale pooling of units-individual units had weak predictive power.  ...  Analogous to the primate visual system, DCNNs develop representations that generalize over image variation, while retaining subject (e.g., gender) and image (e.g., viewpoint) information.  ...  Gender Gender-classification accuracy was measured in the same abbreviated subspaces sampled for the face-identification experiments.  ... 
doi:10.1167/jov.21.8.15 pmid:34379084 pmcid:PMC8363775 fatcat:zskpuoaolze7vjcygt46hntjai

Unravelling the Effect of Image Distortions for Biased Prediction of Pre-trained Face Recognition Models [article]

Puspita Majumdar, Surbhi Mittal, Richa Singh, Mayank Vatsa
2021 arXiv   pre-print
We provide a systematic analysis to evaluate the performance of four state-of-the-art deep face recognition models in the presence of image distortions across different gender and race subgroups.  ...  For the first time, we attempt to answer this question in the context of face recognition.  ...  On varying the resolution of the images from 48 × 48 to 28 × 28, the accuracy of ArcFace drops by 89.74% and 81.93% for subgroup G1 and G2, respectively.  ... 
arXiv:2108.06581v1 fatcat:pjzacpzbfjdp5ikt54joxiod3y

Jointly De-biasing Face Recognition and Demographic Attribute Estimation [article]

Sixue Gong, Xiaoming Liu, Anil K. Jain
2020 arXiv   pre-print
Adversarial learning is adopted to minimize correlation among feature factors so as to abate bias influence from other factors.  ...  The proposed network consists of one identity classifier and three demographic classifiers (for gender, age, and race) that are trained to distinguish identity and demographic attributes, respectively.  ...  Table 2 : 2 Demographic Classification Accuracy (%) by face features.  ... 
arXiv:1911.08080v4 fatcat:2cavhrnfezggjh6jffhpxzhfoy

SensitiveLoss: Improving Accuracy and Fairness of Face Representations with Discrimination-Aware Deep Learning [article]

Ignacio Serna, Aythami Morales, Julian Fierrez, Manuel Cebrian, Nick Obradovich, Iyad Rahwan
2020 arXiv   pre-print
The experiments include tree popular face recognition models and three public databases composed of 64,000 identities from different demographic groups characterized by gender and ethnicity.  ...  We propose a discrimination-aware learning method to improve both accuracy and fairness of biased face recognition algorithms.  ...  Automatic classification algorithms based on these reduced categories show performances of up to 98% accuracy [10] .  ... 
arXiv:2004.11246v2 fatcat:xvccbhv6qjdvhosekaoy32zika

Towards Accuracy-Fairness Paradox: Adversarial Example-based Data Augmentation for Visual Debiasing [article]

Yi Zhang, Jitao Sang
2020 arXiv   pre-print
This paper studies the debiasing problem in the context of image classification tasks.  ...  Results on simulated and real-world debiasing experiments demonstrate the effectiveness of the proposed solution in simultaneously improving model accuracy and fairness.  ...  Specifically, we first trained binary gender classifier д or i with original face images from the CelebA dataset, and then employed I-FGSM [16] to attack each original image to its adversarial image  ... 
arXiv:2007.13632v2 fatcat:3zbxitm6bbbgllkvmk3lfmwfoe

Towards Gender-Neutral Face Descriptors for Mitigating Bias in Face Recognition [article]

Prithviraj Dhar, Joshua Gleason, Hossein Souri, Carlos D. Castillo, Rama Chellappa
2020 arXiv   pre-print
It appears to contribute to gender bias in face recognition, i.e. we find a significant difference in the recognition accuracy of DCNNs on male and female faces.  ...  Therefore, we present a novel 'Adversarial Gender De-biasing algorithm (AGENDA)' to reduce the gender information present in face descriptors obtained from previously trained face recognition networks.  ...  We find that for both Arcface, the Crystalface, the classification accuracy goes down when the face descriptors are transformed using AGENDA or CorrPCA framework, which indicates that gender leakage from  ... 
arXiv:2006.07845v2 fatcat:zrhs2pjzdrfd5kks65a2wydqqa

SensorSift

Miro Enev, Jaeyeon Jung, Liefeng Bo, Xiaofeng Ren, Tadayoshi Kohno
2012 Proceedings of the 28th Annual Computer Security Applications Conference on - ACSAC '12  
To evaluate our approach, we apply SensorSift to the PubFig dataset of celebrity face images, and study how well we can simultaneously hide and reveal various policy combinations of face attributes using  ...  In addition, our sifting transformations led to consistent classification performance when evaluated using a set of five modern machine learning methods (linear SVM, kNearest Neighbors, Random Forests,  ...  Using the annotations provided from the dataset we first cropped the face region from each frame. Next we extracted image features as described in Section 5.  ... 
doi:10.1145/2420950.2420975 dblp:conf/acsac/EnevJBRK12 fatcat:m3swbbkfojepvejnn2qnwibz44

An Efficient Smote-based Model for Dyslexia Prediction

Vani Chakraborty, Research Scholar, Garden City University, Meenatchi Sundaram
2021 International Journal of Information Engineering and Electronic Business  
Also it is observed that the dataset has an unequal distribution of positive and negative cases and so the classification accuracy is compromised if used directly.  ...  It uses a dataset made available from the kaggle repository to predict the risk of dyslexia using various machine learning algorithms.  ...  Since the dataset had unequal distribution of positive and negative values, the classification accuracy of the different machine learning algorithms would be inaccurate.  ... 
doi:10.5815/ijieeb.2021.06.02 fatcat:o6ax4vb6ofd3rfnpntmldz5roe
« Previous Showing results 1 — 15 out of 6,689 results