Maneet Singh

Bio: Maneet Singh is an academic researcher from Indraprastha Institute of Information Technology. The author has contributed to research in topics: Facial recognition system & Autoencoder. The author has an hindex of 15, co-authored 51 publications receiving 605 citations.

MoDest: Multi-module Design Validation for Documents

Abstract: Information extraction (IE) from Visually Rich Documents (VRDs) is a common need for businesses, where extracted information is used for various purposes such as verification, design validation, or compliance. Most of the research in IE from VRDs has focused on textual documents such as invoices and receipts, while extracting information from multi-modal VRDs remains a challenging task. This research presents a novel end-to-end design validation framework for multi-modal VRDs containing textual and visual components, for compliance against a pre-defined set of rules. The proposed Multi-mOdule DESign validaTion (referred to as MoDest) framework constitutes two steps: (i) information extraction using five modules for obtaining the textual and visual components, followed by (ii) validating the extracted components against a pre-defined set of design rules. Given an input multi-modal VRD image, the MoDest framework either accepts or rejects its design while providing an explanation for the decision. The proposed framework is tested for design validation for a particular type of VRDs: banking cards, under the real-world constraint of limited and highly imbalance training data with more than 99% of card designs belonging to one class (accepted). Experimental evaluation on real world images from our in-house dataset demonstrates the effectiveness of the proposed MoDest framework. Analysis drawn from the real-world deployment of the framework further strengthens its utility for design validation.

In-Group Bias in Deep Learning-Based Face Recognition Models Due to Ethnicity and Age

Abstract: Humans are known to favor other individuals who exist in similar groups as them, exhibiting biased behavior, which is termed as in-group bias. The groups could be formed on the basis of ethnicity, age, or even a favorite sports team. Taking cues from aforementioned observation, we inspect if deep learning networks also mimic this human behavior, and are affected by in-group and out-group biases. In this first of its kind research, the behavior of face recognition models is evaluated to understand if similar to humans, models also encode group-specific features for face recognition, along with where bias is encoded in these models. Analysis has been performed for two use-cases of bias: age and ethnicity in face recognition models. Thorough experimental evaluation leads us to several insights: (i) deep learning models focus on different facial regions for different ethnic groups and age groups, and (ii) large variation in face verification performance is also observed across different sub-groups for both known and our own trained deep networks. Based on the observations, a novel bias index is presented for evaluating a trained model’s level of bias. We believe that a better understanding of how deep learning models work and encode bias, along with the proposed bias index would enable researchers to address the challenge of bias in AI, and develop more robust and fairer algorithms for mitigating bias as well as developing fairer models.

Enhancing Fine-Grained Classification for Low Resolution Images

Abstract: Low resolution fine-grained classification has widespread applicability for applications where data is captured at a distance such as surveillance and mobile photography. While fine-grained classification with high resolution images has received significant attention, limited attention has been given to low resolution images. These images suffer from the inherent challenge of limited information content and the absence of fine details useful for sub-category classification. This results in low inter-class variations across samples of visually similar classes. In order to address these challenges, this research proposes a novel attribute-assisted loss, which utilizes ancillary information to learn discriminative features for classification. The proposed loss function enables a model to learn class-specific discriminative features, while incorporating attribute-level separability. Evaluation is performed on multiple datasets with different models, for four resolutions varying from 32x32 to 224x224. Different experiments demonstrate the efficacy of the proposed attributeassisted loss for low resolution fine-grained classification.

Understanding Neural Responses to Face Verification of Cross-Domain Representations

Abstract: Face verification involves identifying whether two faces belong to the same person or not. It relies heavily upon face perception, processing, and the decision making of an individual. This research studies cross-domain face verification, where one face image belongs to a controlled, well-illuminated environment, while the other is of a varying representation having differences in image type or quality. Specifically, two cross-domain face verification tasks are analyzed: controlled-low resolution and controlled-sketch face verification. functional Magnetic Resonance Imaging (fMRI) data has been collected for 23 participants of two ethnic groups while performing face verification. Statistical comparisons were performed with same-domain controlled face verification for both the tasks. Our findings reveal regions of Right Frontal Gyrus, Bilateral Insula, and Right Middle Cingulate Cortex demonstrating higher activation for controlled-sketch face verification, as compared to controlled face verification. Similar analysis were performed for controlled-low resolution face verification, where regions responsible for higher visual load and difficult tasks result in higher activation. Further, stimuli ethnicity differences influence activations for low-resolution face verification but do not affect sketch face verification. Regions of Right Middle Occipital Gyrus and Right Fusiform Gyrus present higher activity, suggesting increased face processing effort for within ethnicity low resolution face verification. We believe the findings of this research will help enable further development in the field of brain-inspired facial recognition algorithms.