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.

Gender and Ethnicity Classification of Iris Images using Deep Class-Encoder

Abstract: Soft biometric modalities have shown their utility in different applications including reducing the search space significantly. This leads to improved recognition performance, reduced computation time, and faster processing of test samples. Some common soft biometric modalities are ethnicity, gender, age, hair color, iris color, presence of facial hair or moles, and markers. This research focuses on performing ethnicity and gender classification on iris images. We present a novel supervised autoencoder based approach, Deep Class-Encoder, which uses class labels to learn discriminative representation for the given sample by mapping the learned feature vector to its label. The proposed model is evaluated on two datasets each for ethnicity and gender classification. The results obtained using the proposed Deep Class-Encoder demonstrate its effectiveness in comparison to existing approaches and state-of-the-art methods.

On Recognizing Face Images With Weight and Age Variations

Abstract: With the increase in age, there are changes in skeletal structure, muscle mass, and body fat. For recognizing faces with age variations, researchers have generally focused on the skeletal structure and muscle mass. However, the effect of change in body fat has not been studied with respect to face recognition. In this paper, we incorporate weight information to improve the performance of face recognition with age variations. The proposed algorithm utilizes neural network and random decision forest to encode age variations across different weight categories. The results are reported on the WhoIsIt database prepared by the authors containing 1109 images from 110 individuals with age and weight variations. The comparison with existing state-of-the-art algorithms and commercial system on WhoIsIt and FG-Net databases shows that the proposed algorithm outperforms existing algorithms significantly.

Transfer Learning Based Evolutionary Algorithm for Composite Face Sketch Recognition

Abstract: Matching facial sketches to digital face images has widespread application in law enforcement scenarios. Recent advancements in technology have led to the availability of sketch generation tools, minimizing the requirement of a sketch artist. While these sketches have helped in manual authentication, matching composite sketches with digital mugshot photos automatically show high modality gap. This research aims to address the task of matching a composite face sketch image to digital images by proposing a transfer learning based evolutionary algorithm. A new feature descriptor, Histogram of Image Moments, has also been presented for encoding features across modalities. Moreover, IIITD Composite Face Sketch Database of 150 subjects is presented to fill the gap due to limited availability of databases in this problem domain. Experimental evaluation and analysis on the proposed dataset show the effectiveness of the transfer learning approach for performing cross-modality recognition.

Dual Directed Capsule Network for Very Low Resolution Image Recognition

Abstract: Very low resolution (VLR) image recognition corresponds to classifying images with resolution 16x16 or less. Though it has widespread applicability when objects are captured at a very large stand-off distance (e.g. surveillance scenario) or from wide angle mobile cameras, it has received limited attention. This research presents a novel Dual Directed Capsule Network model, termed as DirectCapsNet, for addressing VLR digit and face recognition. The proposed architecture utilizes a combination of capsule and convolutional layers for learning an effective VLR recognition model. The architecture also incorporates two novel loss functions: (i) the proposed HR-anchor loss and (ii) the proposed targeted reconstruction loss, in order to overcome the challenges of limited information content in VLR images. The proposed losses use high resolution images as auxiliary data during training to "direct" discriminative feature learning. Multiple experiments for VLR digit classification and VLR face recognition are performed along with comparisons with state-of-the-art algorithms. The proposed DirectCapsNet consistently showcases state-of-the-art results; for example, on the UCCS face database, it shows over 95% face recognition accuracy when 16x16 images are matched with 80x80 images.

Dual Directed Capsule Network for Very Low Resolution Image Recognition

Abstract: Very low resolution (VLR) image recognition corresponds to classifying images with resolution 16x16 or less. Though it has widespread applicability when objects are captured at a very large stand-off distance (e.g. surveillance scenario) or from wide angle mobile cameras, it has received limited attention. This research presents a novel Dual Directed Capsule Network model, termed as DirectCapsNet, for addressing VLR digit and face recognition. The proposed architecture utilizes a combination of capsule and convolutional layers for learning an effective VLR recognition model. The architecture also incorporates two novel loss functions: (i) the proposed HR-anchor loss and (ii) the proposed targeted reconstruction loss, in order to overcome the challenges of limited information content in VLR images. The proposed losses use high resolution images as auxiliary data during training to "direct" discriminative feature learning. Multiple experiments for VLR digit classification and VLR face recognition are performed along with comparisons with state-of-the-art algorithms. The proposed DirectCapsNet consistently showcases state-of-the-art results; for example, on the UCCS face database, it shows over 95\% face recognition accuracy when 16x16 images are matched with 80x80 images.