Computer Vision and Image Processing: 6th International Conference, CVIP 2021, Rupnagar, India, December 3–5, 2021, Revised Selected Papers, Part I ... in Computer and Information Science, 1567)

Preface
Organization
Contents – Part I
Contents – Part II
Classification of Brain Tumor MR Images Using Transfer Learning and Machine Learning Models
	1 Introduction
	2 Literature Review
	3 Working Model
		3.1 Proposed Method-1
		3.2 Proposed Method-2
		3.3 Datasets
		3.4 Performance Metrices
	4 Results and Discussions
	5 Conclusions
	References
Deep-TDRS: An Integrated System for Handwritten Text Detection-Recognition and Conversion to Speech Using Deep Learning
	1 Introduction
	2 Related Works
	3 The Proposed Pipeline
		3.1 Detection System
		3.2 Recognition System
		3.3 Post Processing
		3.4 Text to Speech
	4 Training Details
		4.1 Dataset
		4.2 Implementation
	5 Experimental Results and Analysis
		5.1 Detection
		5.2 Recognition and Post Processing
	6 Conclusion
	References
Computer Aided Diagnosis of Autism Spectrum Disorder Based on Thermal Imaging
	1 Introduction
	2 Related Works
	3 Methodology
		3.1 Study Design and Population
		3.2 Thermal Image Acquisition Procedure
		3.3 Thermal Image Segmentation Algorithm
		3.4 Statistical Feature Extraction
		3.5 Machine Learning and Deep Learning Classifier
		3.6 Statistical Analysis
	4 Results
	5 Conclusion
	References
Efficient High-Resolution Image-to-Image Translation Using Multi-Scale Gradient U-Net
	1 Introduction
	2 Proposed MSG U-Net GAN Model
		2.1 Generator Architecture
		2.2 Discriminator Architecture
		2.3 Loss Function
	3 Experiments and Results
		3.1 Quantitative Results
		3.2 Ablation Study
		3.3 Comparison of MSG U-Net and Pix2Pix Network Complexity and Inference Time
		3.4 Qualitative Results
	4 Conclusion
	References
Generic Multispectral Image Demosaicking Algorithm and New Performance Evaluation Metric
	1 Introduction
	2 Related Work
	3 Proposed Multispectral Demosaicking Algorithm
	4 New Proposed Metric: ANMPSNR
	5 Experimental Results and Discussions
	6 Conclusion and Future Work
	References
A Platform for Large Scale Auto Annotation of Scanned Documents Featuring Real-Time Model Building and Model Pooling
	1 Introduction
	2 System Design
		2.1 Overview of the Platform
		2.2 Design Details of Annotator
	3 Methodology
		3.1 CNN Based Sub-region Specific Models
		3.2 Training Methodology
		3.3 Model Pooling
	4 Data Sets
	5 User Scenario
	6 Results
		6.1 Experimenting with Various Data Sets
		6.2 Tesseract OCR Annotations
	7 Conclusion
	References
AC-CovidNet: Attention Guided Contrastive CNN for Recognition of Covid-19 in Chest X-Ray Images
	1 Introduction
	2 Related Works
	3 Proposed AC-CovidNet CNN Model
		3.1 Background
		3.2 Proposed Model
		3.3 Objective Function
	4 Experimental Setup
		4.1 Dataset Used
		4.2 Training Settings
	5 Experimental Results and Analysis
	6 Conclusion
	References
Application of Deep Learning Techniques for Prostate Cancer Grading Using Histopathological Images
	1 Introduction
	2 Literature Review
	3 Methods
		3.1 Pre-trained CNN Models
		3.2 Pre-trained CNN Models Training
	4 Experiments and Results
		4.1 Dataset
		4.2 Results
	5 Conclusions
	References
Dyadic Interaction Recognition Using Dynamic Representation and Convolutional Neural Network
	1 Introduction
	2 Literature Survey
	3 Proposed Methodology
		3.1 Construction of Dynamic Image
		3.2 Deep Learning Architecture for Action Recognition
	4 Experimental Results and Discussion – Human Interaction Recognition
		4.1 Dataset Details
		4.2 Dynamic Image Generation
		4.3 Recognition Rate with Convolutional Neural Network
	5 Conclusion
	References
Segmentation of Unstructured Scanned Devanagari Newspaper Documents
	1 Introduction
	2 Related Work
	3 Experimental Work
	4 Discussion on Experimental Results
	5 Conclusion
	References
Automatic Classification of Sedimentary Rocks Towards Oil Reservoirs Detection
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Deep Nets for Classification of Different Rocks
		3.2 Robust Rock Segmentation Approach for Pore Space Analysis
	4 Performance Evaluation
		4.1 Ground Truth (GT) Generation
		4.2 Analysis of the Rock Samples Classification Using CNN Models
		4.3 Performance Evaluation of the Segmented Pore Spaces
		4.4 Comparative Analysis with State-of-the-Arts
		4.5 Discussion of Results Significance
	5 Conclusion and Future Work
	References
Signature2Vec - An Algorithm for Reference Frame Agnostic Vectorization of Handwritten Signatures
	1 Introduction
	2 Datasets
	3 Proposed Algorithms
		3.1 Signature Vectorization Algorithm
		3.2 Synthetic Forged Online Signature Generation
		3.3 Online Signature Verification Algorithm
	4 Results and Discussions
		4.1 Scenario I - Assessment of Signature2Vec Embeddings
		4.2 Scenario II - Effectiveness of Signature2Vec in Conjunction with Random Forest
		4.3 Signature2Vec Embeddings - Rotational, Translational and Scale Invariance
	5 Conclusions and Future Directions
	References
Leaf Segmentation and Counting for Phenotyping of Rosette Plants Using Xception-style U-Net and Watershed Algorithm
	1 Introduction
	2 Related Work
	3 Methods
		3.1 Proposed Xception-style U-Net Architecture
		3.2 Marker-Based Watershed Algorithm
		3.3 Dataset
		3.4 Training and Testing of Networks for Leaf Segmentation
		3.5 Leaf Counting Using Watershed Algorithm
	4 Results and Discussion
	5 Conclusion and Future Research Direction
	References
Fast and Secure Video Encryption Using Divide-and-Conquer and Logistic Tent Infinite Collapse Chaotic Map
	1 Introduction
	2 Related Work
	3 Proposed Architectures
		3.1 Pseudo-random Sequence Generation(PRSG)
		3.2 Performance Analysis of Chaotic Sequence
		3.3 Proposed Encryption Scheme
	4 Experimental Results
	5 Conclusion
	References
Visual Localization Using Capsule Networks
	1 Introduction
		1.1 Camera Pose Estimation
		1.2 Literature Review
	2 Method
		2.1 Model Architecture
		2.2 Model Hypothesis
		2.3 Experimentation
	3 Results and Discussions
		3.1 Small NORB
		3.2 Shop Facade
		3.3 Comparison Between PoseCap and PoseNet
		3.4 Feature Visualizations
	4 Future Work and Conclusion
		4.1 Future Work
		4.2 Conclusion
	References
Detection of Cataract from Fundus Images Using Deep Transfer Learning
	1 Introduction
	2 Related Works
	3 Proposed Method
	4 Experimental Results
		4.1 Dataset
		4.2 Data Augmentation
		4.3 Evaluation Metrics
		4.4 Results and Comparison
		4.5 EfficientNetB0
	5 Conclusion
	References
Brain Tumour Segmentation Using Convolution Neural Network
	1 Introduction
	2 Literature Survey
	3 Related Work
	4 Convolution Neural Network for Brain Tumor Segmentation
	5 Network Training Details
	6 Experimental Results
		6.1 Quantitative Evaluation
		6.2 Qualitative Analysis
	7 Conclusion
	References
Signature Based Authentication: A Multi-label Classification Approach to Detect the Language and Forged Sample in Signature
	1 Introduction
	2 Related Work
	3 Proposed Work
		3.1 Finetuning
		3.2 Dataset Preparation
		3.3 Preprocessing
		3.4 Feature Extraction
		3.5 Training of the Artificial Neural Network
	4 Experiment and Results
		4.1 Training Strategy
		4.2 Experimental Setup
		4.3 Dataset
		4.4 Performance Measures
		4.5 Results and Discussion
	5 Conclusion
	References
A Data-Set and a Real-Time Method for Detection of Pointing Gesture from Depth Images
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Creation of a Large Data-Set for Detecting Pointing Gestures
		3.2 Deep Learning Based Detection of Pointing Gesture
	4 Results and Comparison
	5 Conclusion
	References
VISION HELPER: CNN Based Real Time Navigator for the Visually Impaired
	1 Introduction
	2 Related Work
	3 Proposed Model
		3.1 Object Detection
		3.2 Object Location Analysis
		3.3 Auditory Description Generation
	4 Experimental Results
		4.1 Dataset and Training
		4.2 Testing
		4.3 Test Scenarios and Observations
	5 Realtime Application
	6 Conclusion
	References
Structure-Texture Decomposition-Based Enhancement Framework for Weakly Illuminated Images
	1 Introduction
	2 Proposed Methodology
		2.1 TV-based Structure-Texture Image Decomposition
		2.2 Adaptive Luminance Enhancement
		2.3 Texture-Preserved Noise Suppression
		2.4 Reconstruction of Enhanced Image
	3 Experimental Results and Discussion
		3.1 Quantitative Evaluation
	4 Conclusion and Future Work
	References
Low Cost Embedded Vision System for Location and Tracking of a Color Object
	1 Introduction
	2 Capturing and Display of the Image
		2.1 Image Acquisition
		2.2 Displaying the Image
	3 Image Processing
		3.1 Image Segmentation
		3.2 Description of the Region of Interest
		3.3 Tracking
	4 Tests and Results
		4.1 Image Segmentation
		4.2 Detection of Distinct Color Objects
		4.3 Location and Tracking
		4.4 Cost and Power Consumption
	5 Conclusions
	References
Towards Label-Free Few-Shot Learning: How Far Can We Go?
	1 Introduction
	2 Related Work
		2.1 Related Perspectives
	3 Approach: Few-Shot Learning with Almost No Labels
		3.1 Testing Framework
	4 Experiments
		4.1 Experimental Setup
		4.2 Results
		4.3 Ablation Studies
	5 Conclusion
	References
AB-net: Adult- Baby Net
	1 Introduction
	2 Approach
		2.1 Data Preparation Methodology
		2.2 Training
		2.3 Cascade 1:
	3 Conclusion
	References
Polarimetric SAR Classification: Fast Learning with k-Maximum Likelihood Estimator
	1 Introduction
	2 Polarimetric SAR Data Representation
	3 Methodology
	4 Experimental Results
	5 Conclusion
	References
Leveraging Discriminative Cues for Masked Face Recognition in Post COVID World
	1 Introduction
	2 Related Work
	3 Proposed Work
		3.1 Single Shot Detection
		3.2 Facial Landmark Localization with Mask Recognition
		3.3 Mask the Dataset
		3.4 Mask Removal
		3.5 Embedding Generation
	4 Results
	5 Conclusion
	References
Pretreatment Identification of Oral Leukoplakia and Oral Erythroplakia Metastasis Using Deep Learning Neural Networks
	1 Introduction
		1.1 Oral Cancer Types
	2 Literature Review
	3 Proposed Work
		3.1 Data Preparation
		3.2 Model Training
		3.3 Results and Discussion
	4 Conclusion
	References
Soft Biometric Based Person Retrieval for Burglary Investigation
	1 Introduction
	2 Related Work
		2.1 Person Retrieval Methodologies
		2.2 Semantic Query Based Person Retrieval
	3 The Proposed Soft-Biomatric Based Person Retrieval Framework
		3.1 Height Estimation
		3.2 Attire Colour Classification Module
		3.3 Gender Classification
	4 Experiment
		4.1 Dataset Overview and Performance Metric
		4.2 Implementation Details
	5 Experimental Evaluation and Discussion
	6 Conclusion and Future Work
	References
A Deep Learning Framework for the Classification of Lung Diseases Using Chest X-Ray Images
	1 Introduction
	2 Materials and Methods
		2.1 Data Set
		2.2 Methodology
	3 Results and Discussion
		3.1 Evaluation
		3.2 Experiments
		3.3 Discussion
	4 Conclusion
	References
Scene Graph Generation with Geometric Context
	1 Introduction
	2 Related Work
		2.1 Visual Relationship Detection
		2.2 Graph Inference
		2.3 Routing Network with Embedded Knowledge
	3 Proposed Solution
		3.1 Geometric Context
	4 Dataset
		4.1 Visual Genome
	5 Implementation
		5.1 Tasks
		5.2 Evaluation Metrics
	6 Results
	7 Conclusion
	References
Deep Color Spaces for Fingerphoto Presentation Attack Detection in Mobile Devices
	1 Introduction
	2 Related Work
	3 The Proposed System
		3.1 Architecture and Fine-Tuning of the DeepNets
		3.2 Implementation of Mobile Deep Learning
	4 Experimental Results
		4.1 Dataset
		4.2 Evaluation Procedure
		4.3 Results
	5 Conclusions
	References
Cancelable Template Generation Using Convolutional Autoencoder and RandNet
	1 Introduction
	2 Literature Review
	3 Methodology
		3.1 Proposed CAE
		3.2 Rank Based Partition Network
		3.3 RandNet and Random Permutation Flip Mechanism
		3.4 Template Generation
	4 Results and Discussion
		4.1 Performance Analysis
		4.2 Unlinkability Analysis
		4.3 Non-invertibility Analysis
	5 Conclusion
	References
Homogeneous and Non-homogeneous Image Dehazing Using Deep Neural Network
	1 Introduction
	2 Related Works
		2.1 Traditional Techniques
		2.2 Deep Learning Models
	3 Proposed Model
		3.1 Base Model
		3.2 Adapting Base Model for Visual Enhancement
	4 Experimental Results
		4.1 Datasets and Model Details
		4.2 Quantitative Evaluation
		4.3 Qualitative Results
	5 Conclusion
	References
Improved Periocular Recognition Through Blend of Handcrafted and Deep Features
	1 Introduction
	2 Literature Survey
	3 Proposed Methodology
		3.1 Handcrafted Feature Extraction
		3.2 Deep Feature Extraction
		3.3 Fusion
	4 Experiments and Results
		4.1 Database
		4.2 Results and Discussion
	5 Conclusion
	References
Kernels for Incoherent Projection and Orthogonal Matching Pursuit
	1 Introduction
	2 Basics of Compressed Sensing and Kernels
		2.1 Compressed Sensing
	3 Motivation for Present Work and Contribution
		3.1 Kernel Trick
	4 Kernel OMP
		4.1 Theory
	5 Classification of ECG Signals via OMP
	6 Extraction of Beats and Experimental Setup
		6.1 Extraction of Beats
	7 Simulation Work and Conclusion
	References
AAUNet: An Attention Augmented Convolution Based UNet for Change Detection in High Resolution Satellite Images
	1 Introduction
	2 Related Work
	3 Proposed Framework
		3.1 Dataset Description
		3.2 Implementation Details
		3.3 UNet Structure Based on Encoder-Decoder for Semantic Segmentation
		3.4 Self-attention Mechanisms in Architecture
		3.5 Attention Augmented Convolution in Network
		3.6 Attention Augmented Convolution over Image
		3.7 Depthwise Separable Convolution
		3.8 Proposed Attention Augmented Convolution Based UNet Architecture
		3.9 Training Parameters
	4 Results and Discussion
	5 Conclusion
	References
A Large Volume Natural Tamil Character Dataset
	1 Introduction
	2 Materials and Methods
		2.1 Natural Image Tamil Character Dataset
		2.2 Feature Extraction Methods
	3 Results and Discussion
	4 Conclusion
	References
Datasets of Wireless Capsule Endoscopy for AI-Enabled Techniques
	1 Introduction
	2 Open WCE Datasets
		2.1 Kvasir-Capsule ch384
		2.2 Red Lesion Endoscopy Dataset ch385
		2.3 Annotated Bleeding Dataset ch386
		2.4 EndoSLAM ch387
		2.5 KID Project ch388
	3 Private WCE Datasets
		3.1 CAD-CAP ch389
		3.2 Lesion Detection Dataset ch38il
		3.3 Crohn IPI ch38val
		3.4 Cleansing Small Bowel Capsule Endoscopy ch38mal
		3.5 Bleeding Dataset ch38la
		3.6 WCE-2019-Video ch38lan
		3.7 Celiac Disease WCE Dataset ch38ma
		3.8 AIIMS WCE Dataset
	4 Miscellaneous
	5 Conclusion
	References
Moving Objects Detection in Intricate Scenes via Spatio-Temporal Co-occurrence Based Background Subtraction
	1 Introduction
	2 Proposed System Framework
		2.1 Multi-frame Spatio-Temporal Co-occurrence (MSC) Descriptor
		2.2 Feature Vector Generation
		2.3 Background Modelling
		2.4 Foreground Detection
	3 Experimental Results
	4 Conclusion
	References
Script Identification in Natural Scene Text Images by Learning Local and Global Features on Inception Net
	1 Introduction
		1.1 The Past Work
		1.2 The Present Work
		1.3 Contribution
	2 Methodology
	3 Experiments
	4 Conclusion and Future Scope
	References
Survey of Leukemia Cancer Cell Detection Using Image Processing
	1 Introduction
	2 Literature Review
		2.1 Conventional Methods
		2.2 Machine Learning Methods
		2.3 Deep Learning Methods
	3 Datasets
	4 Evaluation Model
	5 Research Problem
	6 Challenges in Cell Segmentation and Classification
	7 Ideas for Further Research
	8 Analysis
	9 Conclusion
	References
MS-Net: A CNN Architecture for Agriculture Pattern Segmentation in Aerial Images
	1 Introduction
	2 Related Work
	3 Materials and Methods
		3.1 Dataset
	4 Proposed Method
		4.1 MS-Net Architecture
	5 Experimental Results
		5.1 Evaluation Metric
		5.2 Results and Discussion
	6 Conclusion
	References
Vision Transformer for Plant Disease Detection: PlantViT
	1 Introduction
	2 Related Work
	3 Plant Vision Transformer: PlantViT
	4 Result and Discussion
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 Results
		4.4 Discussion
	5 Conclusion and Future Work
	References
Evaluation of Detection and Segmentation Tasks on Driving Datasets
	1 Introduction
	2 Related Work
	3 Experiments
		3.1 Datasets
		3.2 Setup
		3.3 Performance Metric
	4 Results
		4.1 Object Detection
		4.2 Semantic Segmentation
		4.3 Instance Segmentation
	5 Summary
	References
Classification of Gender in Celebrity Cartoon Images
	1 Introduction
		1.1 Related Works
	2 Proposed Model
		2.1 Pre-processing
		2.2 Feature Extraction
		2.3 Classifiers
		2.4 Integrated Model
	3 Experiment Results
		3.1 Dataset
		3.2 Experimental Setup
		3.3 Experimental Results
	4 Comparative Analyses
	5 Conclusion
	References
Localization of Polyps in WCE Images Using Deep Learning Segmentation Methods: A Comparative Study
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 FCN
		3.2 SegNet
		3.3 UNet
		3.4 PSPNet
	4 Results
	5 Conclusion
	References
Evaluation of Deep Architectures for Facial Emotion Recognition
	1 Introduction
	2 Deep Learning Architectures
		2.1 Visual Geometry Group (VGG)
		2.2 Pyramid Neural Network (PyramidNet)
		2.3 Extended Residual Neural Network (ResNeXt)
	3 Methodology
		3.1 Data Collection
		3.2 Data Preprocessing
		3.3 Model Creation
		3.4 Emotion Detection
	4 Experimental Result
	5 Conclusion
	References
Adaptive Rough-Fuzzy Kernelized Clustering Algorithm for Noisy Brain MRI Tissue Segmentation
	1 Introduction
	2 Ground Work
		2.1 Fuzzy Set Theory
		2.2 Fuzzy C Means Clustering (FCM)
		2.3 Kernel Methods
		2.4 Rough Set Theory
	3 Proposed Methodology
		3.1 Adaptive Kernelized Distance Metric
		3.2 Objective Function
		3.3 Membership Function and Cluster Centroid Computation
	4 Performance Comparison and Results
	5 Conclusion
	References
On the Prospects of Latent MasterPrints
	1 Introduction
	2 Latent Fingerprint Identification
	3 MasterPrint Vulnerability
	4 Experimental Setup
	5 Minutiae Detection and Minutiae Matching
	6 Result Analysis
	7 Future Work
	8 Conclusion
	References
Author Index