Computer Vision and Image Processing: 6th International Conference, CVIP 2021, Rupnagar, India, December 3–5, 2021, Revised Selected Papers, Part II ... in Computer and Information Science, 1568)

Preface
Organization
Contents – Part II
Contents – Part I
Handwritten Text Retrieval from Unlabeled Collections
	1 Introduction
	2 Related Works
	3 Proposed Framework
	4 Experiments and Results
		4.1 Training Phase
		4.2 Evaluation Datasets
		4.3 Evaluation and Discussion
		4.4 Comparative Results for Retrieval
	5 Conclusion
	References
Detecting Document Forgery Using Hyperspectral Imaging and Machine Learning
	1 Introduction
	2 Literature Survey
	3 Dataset Description
	4 Methodology
		4.1 Preprocessing
		4.2 Dimension Reduction
		4.3 Machine Learning
	5 Results
		5.1 Machine Learning Without Dimension Reduction
		5.2 Machine Learning with Dimension Reduction
		5.3 Comparison with Existing Methods
	6 Conclusions and Future Work
	References
An Hour-Glass CNN for Language Identification of Indic Texts in Digital Images
	1 Introduction
	2 Related Works
	3 Proposed Model and Dataset
	4 Experiments and Results
		4.1 Training and Testing
		4.2 Analysis of Results
	5 Conclusion
	References
Single Frame-Based Video Dehazing with Adversarial Learning
	1 Introduction
	2 Literature Review
		2.1 Prior Based
		2.2 Multi-image Fusion
		2.3 Learning-Based
	3 Proposed Method
		3.1 Dilated Residual Block (DRB)
		3.2 Skip Connection
		3.3 Loss Function
	4 Experimental Results
		4.1 Quantitative Analysis
		4.2 Qualitative Analysis
	5 Conclusion
	References
Spatio-Temporal Event Segmentation for Wildlife Extended Videos
	1 Introduction
	2 Relevant Work
	3 Methodology
		3.1 Input Encoding
		3.2 Attention Unit
		3.3 Future Prediction Layer
		3.4 Loss Function
		3.5 Error Gate
	4 Experimental Evaluation
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 Ablative Studies
		4.4 Quantitative Evaluation
		4.5 Qualitative Evaluation
	5 Conclusion
	References
Comparative Analysis of Machine Learning and Deep Learning Models for Ship Classification from Satellite Images
	1 Introduction
	2 Literature Review
	3 Methodology
		3.1 The Dataset
		3.2 Pre-processing
		3.3 Machine Learning Approach
		3.4 Deep Learning Approach
	4 Results and Discussions
	5 Conclusion
	References
Channel Difference Based Regeneration Architecture for Fake Colorized Image Detection
	1 Introduction
	2 Literature Survey
		2.1 Contributions of the Work
	3 Proposed Framework
		3.1 Proposed Channel Difference Map
		3.2 Image Regeneration Network
		3.3 Fake Colorized Image Detection Network (DCDNet)
	4 Datasets and Evaluation Metric
	5 Training of the Proposed Networks
	6 Results and Discussion
		6.1 Result Analysis
		6.2 Ablation Study
	7 Conclusion
	References
DenseASPP Enriched Residual Network Towards Visual Saliency Prediction
	1 Introduction
	2 Related Work
	3 Proposed Model
	4 Experimental Analysis
		4.1 Saliency Prediction Datasets
		4.2 Implementation Details
		4.3 Comparison Results
	5 Conclusion
	References
Brain MRI and CT Image Fusion Using Generative Adversarial Network
	1 Introduction
	2 Related Works
	3 Proposed Method
	4 Results and Discussions
	5 Conclusion
	References
MFCA-Net: Multiscale Feature Fusion with Channel-Wise Attention Network for Automatic Liver Segmentation from CT Images
	1 Introduction
	2 Related Literature
	3 Methodology
		3.1 Multiscale Feature Representation
	4 Experimental Settings and Result Analysis
		4.1 Dataset and Preprocessing
		4.2 Training Settings and Performance Metrics
		4.3 Result Analysis and Discussion
	5 Conclusion
	References
Automatic Double Contact Fault Detection in Outdoor Volleyball Videos
	1 Introduction
		1.1 What is a Double Contact Fault?
		1.2 Our Solution
		1.3 Contributions
	2 Literature Review
	3 Proposed Work
		3.1 Detect and Track the Ball
		3.2 Find the Contact Point
		3.3 Deep Feature-Based Classifier
	4 Experiments and Results
		4.1 Dataset
		4.2 Results of Ball Tracking and Detection
		4.3 Results for the Double Contact Fault Classifier
	5 Conclusions
	References
Classroom Slide Narration System
	1 Introduction
	2 Related Work
	3 Classroom Slide Narration System
		3.1 Slide Segmentation Module
		3.2 Information Extraction Module
		3.3 Audio Creation Module
	4 Experiments
		4.1 Experimental Setup
		4.2 Dataset
		4.3 Ablation Study
		4.4 Comparison with State-of-the-Art Techniques
		4.5 Narration System Evaluation
	5 Summary
	References
Humanoid Robot - Spark
	1 Introduction
	2 Block Diagram
		2.1 Spark Face
		2.2 Spark Torso and Arm
		2.3 Spark Gimbal
		2.4 Spark Legs
	3 Methodology
		3.1 Hardware Architecture
		3.2 Features
		3.3 Software Architecture
		3.4 Balancing Mechanism
		3.5 Walking Mechanism
	4 Results
		4.1 Image Processing Results
		4.2 Balancing Results
		4.3 Gesture Results
	5 Applications
	6 Conclusion
	References
Attention-Based Deep Autoencoder for Hyperspectral Image Denoising
	1 Introduction
	2 Image Formation Model
	3 Proposed Methodology
		3.1 Layer-Wise Attention Block
		3.2 Layer-Wise Skip Connection Block
		3.3 Training Details
	4 Experimental Results
	5 Conclusion and Future Work
	References
Feature Modulating Two-Stream Deep Convolutional Neural Network for Glaucoma Detection in Fundus Images
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Feature Extraction
		3.2 Feature Modulation
		3.3 Classification
	4 Experimental Results and Analysis
		4.1 Dataset Used
		4.2 Implementation Details and Performance Measures
		4.3 Ablation Study
		4.4 State-of-the-Art Performance Comparison
	5 Conclusion
	References
Retinal Image Quality Assessment Using Sharpness and Connected Components
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Overview
		3.2 Feature Extraction
	4 Results and Discussion
		4.1 Dataset Description
		4.2 Performance Analysis on Public Dataset
		4.3 Performance Analysis on Private Dataset
	5 Conclusion
	References
(MS)2EDNet: Multiscale Motion Saliency Deep Network for Moving Object Detection
	1 Introduction
	2 Related Work
	3 Proposed Network Framework
		3.1 Motion-Saliency Network(MSNet)
		3.2 Background Estimation (BE)
		3.3 Background Estimation Using CNN (BENet)
		3.4 SaliencyEstimation Network (SENet)
		3.5 Multiscale Encoder-Decoder Network (MSEDNet)
		3.6 Computational Complexity of MSEDNet
	4 Experimental Results and Discussions
		4.1 Results Analysis on CDnet-2014 Dataset
		4.2 Results Analysis on Wallflower Database
	5 Conclusions
	References
Video Enhancement with Single Frame
	1 Introduction
	2 Literature Review
		2.1 Prior Based
		2.2 Multi-image Fusion
		2.3 Learning-Based
	3 Proposed Method
		3.1 Encoder Block (EB)
		3.2 Decoder Block (DB)
		3.3 Loss Function
	4 Experimental Results
		4.1 Quantitative Analysis
		4.2 Qualitative Analysis
	5 Conclusion
	References
Blind Video Quality Assessment Using Fusion of Novel Structural Features and Deep Features
	1 Introduction
	2 Proposed Blind VQA Model
		2.1 Structural Feature Extraction
		2.2 Deep Semantic Feature Extraction
		2.3 Regression Scheme
		2.4 Proposed Methodology
	3 Experimental Results
		3.1 VQA Datasets
		3.2 Experiments
		3.3 Performance Analysis
	4 Conclusion
	References
On-Device Spatial Attention Based Sequence Learning Approach for Scene Text Script Identification
	1 Introduction
	2 Methodology
		2.1 Spatial Attention Based Feature Extractor
		2.2 Sequence Modelling
		2.3 Prediction
	3 Experiments
		3.1 Dataset
		3.2 Implementation Details
		3.3 Effectiveness of Spatial Attention and Residue Convolutional Blocks for Script Identification
		3.4 Comparison with State-of-the-Art Script Identification Methods
	4 Conclusion
	References
Post-harvest Handling of Mangoes: An Integrated Solution Using Machine Learning Approach
	1 Introduction
	2 Proposed Model for Grading
		2.1 Pre-processing/Segmentation
		2.2 Feature Extraction/Selection
		2.3 Classification
		2.4 Internal Defect Detection
	3 Experimentation
		3.1 Dataset
		3.2 Proposed Model in This Study
	4 Further Works
	5 Conclusion
	References
Morphological Gradient Analysis and Contour Feature Learning for Locating Text in Natural Scene Images
	1 Introduction
	2 Related Work
	3 Dataset Description
	4 Proposed Method
		4.1 Phase 1: Training
		4.2 Phase 2: Testing
	5 Experimental Results
	6 Conclusion
	References
Introspecting Local Binary Feature Vectors for Classification of Radiographic Weld Images
	1 Introduction
	2 Database for Present Research Work
	3 Local Binary Feature Vectors
		3.1 Local Binary Pattern (LBP)
		3.2 Uniform Local Binary Pattern (LBPu2)
		3.3 Rotational Invariant Local Binary Pattern (LBPri)
		3.4 Local Binary Pattern Histogram Fourier Features (LBP-HF)
		3.5 Adaptive Local Binary Pattern (ALBP)
		3.6 Complete Local Binary Pattern (CLBP)
	4 Feature Reduction by PCA
	5 Methodology
		5.1 Assessment of Different Texture Feature Extraction Techniques
	6 Results and Discussion
		6.1 Performance Evaluation of Feature Extraction Techniques
		6.2 Performance Evaluation Using ANN
	7 Conclusion
	References
Performance Evaluation of Deep Learning Models for Ship Detection
	1 Introduction
	2 Related Work
	3 Training Data
	4 Methods
	5 Experiments
	6 Evaluation and Results
	7 Conclusion
	References
COVID-19 Social Distance Surveillance Using Deep Learning
	1 Introduction
	2 Methodology
		2.1 Human Detection
		2.2 Top-Down View Transformation
		2.3 Distance Calculation
		2.4 Pixel Depth-Based Distance Thresholding and Social Distance Evaluation
	3 Dataset and Evaluation Metrics
	4 Results and Discussion
	5 Conclusion
	References
Towards Semi-supervised Tree Canopy Detection and Extraction from UAV Images
	1 Introduction
	2 Methodology and Data
		2.1 Methodology
		2.2 Data
	3 Results and Discussion
	4 Conclusion
	References
Pose Guided Controllable Gesture to Gesture Translation
	1 Introduction
	2 Review of Literature
	3 Proposed Method
	4 Experimental Set-Up
		4.1 Datasets
		4.2 Results
	5 Conclusion
	References
EDR: Enriched Deep Residual Framework with Image Reconstruction for Medical Image Retrieval
	1 Introduction
	2 Proposed EDR: Enriched Deep Residual Framework For MCBIR
		2.1 Enriched Deep Residual Framework For Medical Image Reconstruction
		2.2 Effective Feature Extraction and Similarity Indexing
	3 Experimental Results and Discussion
		3.1 Retrieval Accuracy on ILD Database
		3.2 Retrieval Accuracy on VIA/I-ELCAP-CT Dataset
	4 Conclusion
	References
Depth Estimation Using Sparse Depth and Transformer
	1 Introduction
	2 Literature Survey
	3 Proposed Method
		3.1 Sparse Depth
		3.2 Proposed Network
		3.3 Losses
	4 Experimental Analysis
		4.1 Training Details
		4.2 Dataset Description
	5 Results
	6 Conclusion
	References
Analysis of Loss Functions for Image Reconstruction Using Convolutional Autoencoder
	1 Introduction
	2 Related Work
	3 Convolutional Autoencoder (CAE)
	4 Loss Functions
	5 Experimental Setup
		5.1 Datasets Used for Experimentation and Pre-processing
		5.2 Evaluation Metrics
	6 Results and Discussion
	7 Conclusion
	References
Fuzzy Entropy k-Plane Clustering Method and Its Application to Medical Image Segmentation
	1 Introduction
	2 Preliminaries and Related Work
		2.1 Concept of Entropy for Fuzzy Sets
		2.2 FEC
		2.3 kPC
	3 Proposed Fuzzy Entropy k-Plane Clustering Approach
	4 Dataset and Experimental Results
		4.1 Datasets
		4.2 Evaluation Metrics
		4.3 Results on Simulated Brainweb MRI Dataset
		4.4 Results on Real IBSR Brain MRI Dataset
		4.5 Results on Real MRBrainS18 MRI Dataset
	5 Conclusion and Future Works
	References
FMD-cGAN: Fast Motion Deblurring Using Conditional Generative Adversarial Networks
	1 Introduction
	2 Background
		2.1 Image Deblurring
		2.2 Generative Adversarial Networks
	3 Related Works
	4 Our Method
		4.1 Network Architecture
		4.2 Loss Functions
		4.3 Training Datasets
	5 Training Details
	6 Experimental Results
		6.1 Quantitative Evaluation on GoPro Dataset
		6.2 Quantitative Evaluation on REDS Dataset
		6.3 Visual Comparison
	7 Ablation Study
	8 Conclusion
	References
Region Extraction Based Approach for Cigarette Usage Classification Using Deep Learning
	1 Introduction
	2 Related Work
	3 Proposed Methodology
		3.1 Region Extraction Module
		3.2 Classification Module
		3.3 Detection Module
	4 Experiments and Results
		4.1 Implementation
		4.2 Results and Evaluation
		4.3 Comparison with State-of-the-Art Approaches
		4.4 Discussion
	5 Conclusion and Future Work
	References
Fire Detection Model Using Deep Learning Techniques
	1 Introduction
	2 Literature Review
	3 Methodology
		3.1 About Dataset
		3.2 Transfer Learning
		3.3 Optical Flow
	4 Testing and Analysis
		4.1 Overall Performance of System on Fire Detection
		4.2 Comparison of Proposed Method with Recent Papers
	5 Conclusion
	References
Two Novel Methods for Multiple Kinect v2 Sensor Calibration
	1 Related Works
	2 Theoretical Background
	3 Method I – Calibration Procedure for Pair-wise Eight Cameras
		3.1 Merged Point Clouds (Ensuring Successful Calibration)
	4 Method II – Procedure for Simultaneous Camera Calibration
	5 Result Analysis and Discussion
	6 Comparison with Existing Methods
	7 Conclusion and Future Work
	References
Residual Inception Cycle-Consistent Adversarial Networks
	1 Introduction
	2 Literature Review
	3 Proposed Method
		3.1 Residual-Inception Module
		3.2 Cyclic Perceptual-Consistency Loss
	4 Experiments and Results
		4.1 Results on I-haze Dataset
		4.2 Results on Rain1200 Dataset
		4.3 Results on Horse2zebra Dataset
	5 Conclusion
	References
MAG-Net: A Memory Augmented Generative Framework for Video Anomaly Detection Using Extrapolation
	1 Introduction
	2 Literature Review
	3 Proposed Method: Memory Augmented Generative Network
		3.1 Channel Attention (CA)
		3.2 Pixel Attention (PA)
		3.3 Memory Module
	4 Experimental Setup
	5 Results and Discussion
	6 Conclusion
	References
Hand Gesture Recognition Using CBAM-RetinaNet
	1 Introduction
	2 Proposed Network
		2.1 ResNet-18 Backbone
		2.2 Convolutional Block Attention Module
		2.3 Feature Pyramid Network
		2.4 Classification Subnetwork
		2.5 Box Regression Subnetwork
	3 Experimentation
		3.1 Dataset
		3.2 Position of CBAM in the Network
	4 Results
	5 HCI Application
	6 Conclusion
	References
Elderly Patient Fall Detection Using Video Surveillance
	1 Introduction
	2 Related Work
	3 Proposed Approach
		3.1 Dataset
		3.2 Surveillance Video
		3.3 RGB Frames
		3.4 Stream 1
		3.5 Stream 2
		3.6 Results
		3.7 Conclusion and Future Work
	References
OGGN: A Novel Generalized Oracle Guided Generative Architecture for Modelling Inverse Function of Artificial Neural Networks
	1 Introduction
	2 Related Work
	3 Proposed Oracle Guided Generative Neural Network
		3.1 Constraint Functions
	4 Experiments and Results
		4.1 Dataset Preparation
		4.2 Experimental Settings
		4.3 Use Cases of the Proposed OGGN Architecture
	5 Conclusions and Future Work
	References
Deep Learning Based DR Medical Image Classification
	1 Introduction
	2 Literature Survey
	3 Proposed Framework
		3.1 Image Reconstruction
		3.2 Image Classification
	4 Result Analysis
	5 Conclusion
	References
Human Action Recognition in Still Images
	1 Introduction
		1.1 Human Action Recognition
		1.2 Human Action Recognition in Still Images
	2 Related Work
	3 Proposed Approach
		3.1 Dataset Used
		3.2 Methodology
	4 Results
	5 Conclusion and Future Work
	References
Enhancing Unsupervised Video Representation Learning by Temporal Contrastive Modelling Using 2D CNN
	1 Introduction
	2 Literature Review
	3 Proposed Method
		3.1 Sequence Order Prediction
		3.2 Content-Aware Contrastive Learning
		3.3 Implementation
	4 Experimental Settings
	5 Results
		5.1 Influence of Joint Learning on Retrieval Performance
		5.2 Comparison to Existing Methods in Context of Retrieval Performance
		5.3 Action Recognition
	6 Conclusion
	References
Deep Two-Stage LiDAR Depth Completion
	1 Introduction
	2 Literature Survey
		2.1 LiDAR Depth Completion
	3 Residual Dense Blocks
	4 Iterative Propagation
	5 Design Overview
	6 Loss Function
	7 Experiments
		7.1 Dataset and Training Details
		7.2 Evaluation Metrics
		7.3 Comparison with the State-of-the-Art
		7.4 Ablation Experiments
	8 Conclusion
	References
3D Multi-voxel Pattern Based Machine Learning for Multi-center fMRI Data Normalization
	1 Introduction
	2 Materials and Methods
		2.1 Dataset
		2.2 Data Analysis
		2.3 Machine Learning
	3 Experimental Results
		3.1 Multiple Subjects - Multiple Centers (Same Field Strength)
		3.2 Multiple Subjects - Multiple Centers (Different Field Strengths)
	4 Discussion
	5 Conclusion
	References
Efficient Approximation of Curve-Shaped Objects in Z2 Based on the Maximum Difference Between Discrete Curvature Values
	1 Introduction
		1.1 Existing Approximation Methods
		1.2 Discrete Curvature Estimation
	2 Cubic Curve Approximation
		2.1 Proposed Algorithm
	3 Observations and Results
	4 Conclusion
	References
Exploring the Role of Adversarial Attacks in Image Anti-forensics
	1 Introduction
	2 Proposed Anti-forensic Framework Based on Adversarial Attacks
		2.1 Fast Gradient Sign Method
		2.2 Carlini & Wagner Attack
		2.3 Projected Gradient Descent
	3 Experiment Results
	4 Conclusions
	References
A Novel Artificial Intelligence-Based Lung Nodule Segmentation and Classification System on CT Scans
	1 Introduction
	2 Materials and Methods
		2.1 Materials
		2.2 Methods
	3 Results and Discussion
		3.1 Nodule Segmentation
		3.2 Nodule Classification
	4 Conclusion
	References
Author Index