10993nam 2200541 450 99646439300331620211015232907.0981-16-1092-4(CKB)4100000011807087(MiAaPQ)EBC6531637(Au-PeEL)EBL6531637(OCoLC)1244630978(PPN)254720528(EXLCZ)99410000001180708720211015d2021 uy 0engurcnu||||||||txtrdacontentcrdamediacrrdacarrierComputer vision and image processing 5th international conference, CVIP 2020, Prayagraj, India, October 16-18, 2020, revised selected papers, part I /edited by Satish Kumar Singh [and three others]Singapore :Springer,[2021]©20211 online resource (571 pages)Communications in Computer and Information Science ;1377981-16-1091-6 Intro -- Preface -- Organization -- Contents - Part II -- A Comparative Analysis on AI Techniques for Grape Leaf Disease Recognition -- 1 Introduction -- 2 Terminologies in Grape Leaf Disease -- 2.1 Black Rot -- 2.2 Leaf Blight -- 2.3 Mildew -- 2.4 Downy Mildew -- 2.5 Grapevine Measles -- 2.6 Anthracnose -- 2.7 Grey Mold -- 3 Survey on Grape Leaf Disease Recognition and Systematization -- 3.1 Techniques Based on Machine Learning -- 3.2 Techniques Based on Deep Learning -- 4 Results and Discussions -- 5 Conclusion and Future Works -- References -- Sign Language Recognition Using Cluster and Chunk-Based Feature Extraction and Symbolic Representation -- Abstract -- 1 Introduction -- 2 Related Work -- 3 The Proposed Methodology -- 3.1 Face and Hand Segmentation -- 3.2 Feature Extraction -- 3.3 Sign Representation -- 3.4 Symbolic Representation of the Signs -- 4 Experimentation and Validation -- 5 Conclusion -- Acknowledgment -- References -- Action Recognition in Haze Using an Efficient Fusion of Spatial and Temporal Features -- 1 Introduction -- 2 Background and Related Works -- 2.1 Dehazing a Video -- 2.2 Human Action Recognition -- 2.3 Background -- 3 Proposed Method -- 3.1 Dehazing the Frames -- 3.2 Extracting Spatial Features -- 3.3 Bidirectional LSTM (DB-LSTM) for Temporal Feature -- 4 Experiments and Results -- 4.1 Hazy Dataset -- 4.2 Results and Discussions -- 5 Summary and Future Work -- References -- Human Action Recognition from 3D Landmark Points of the Performer -- 1 Introduction and Related Works -- 2 Proposed Method -- 2.1 Extraction of 3D Landmark Points -- 2.2 Classification of Actions Using 3D Landmark Points -- 3 Dataset and Experiments -- 3.1 Dataset -- 3.2 Experimental Set up -- 4 Results -- 5 Conclusions -- References -- A Combined Wavelet and Variational Mode Decomposition Approach for Denoising Texture Images -- 1 Introduction.2 Review -- 2.1 Variational Mode Decomposition -- 2.2 Wavelet Transform -- 3 Proposed Methods -- 4 Experimental Results -- 4.1 Wavelet -- 4.2 VMD -- 4.3 VMD-WT -- 4.4 Proposed WT-VMD -- 5 Conclusion -- References -- Two-Image Approach to Reflection Removal with Deep Learning -- Abstract -- 1 Introduction -- 2 Related Work -- 2.1 Non-learning Based -- 2.2 Learning Based -- 3 Proposed Methodology -- 3.1 Dataset Creation -- 3.2 Network Parameters -- 3.2.1 Structure -- 3.2.2 Loss Function -- 4 Results -- 5 Discussion -- 6 Limitation and Future Work -- 7 Conclusion -- References -- Visual Question Answering Using Deep Learning: A Survey and Performance Analysis -- 1 Introduction -- 2 Datasets -- 3 Deep Learning Based VQA Methods -- 4 Experimental Results and Analysis -- 5 Conclusion -- References -- Image Aesthetic Assessment: A Deep Learning Approach Using Class Activation Map -- 1 Introduction -- 2 Deep CNN Models for IAA -- 3 Image Pre-processing and Two Channel CNN -- 3.1 Image Pre-processing -- 3.2 Two Channel CNN for IAA -- 4 Experiments -- 4.1 Comparison of Different Architectures -- 4.2 Comparison with Various Approaches -- 5 Conclusion -- References -- RingFIR: A Large Volume Earring Dataset for Fashion Image Retrieval -- 1 Introduction -- 2 Related Works -- 3 Proposed Dataset and Benchmark -- 4 Benchmarking Methods and Discussion -- 5 Conclusion -- References -- Feature Selection and Feature Manifold for Age Estimation -- 1 Introduction -- 2 Related Work -- 2.1 Aging Feature -- 2.2 Age Regression -- 3 Proposed Work -- 3.1 Aging Manifold Features -- 3.2 Feature Selection -- 3.3 Regression -- 4 Experiments and Results -- 4.1 Experimental Setup -- 4.2 Experiments and Results -- 5 Conclusion -- References -- Degraded Document Image Binarization Using Active Contour Model -- 1 Introduction -- 2 Proposed Method -- 2.1 Pre-processing.2.2 Initial Mask Calculation -- 2.3 Active Contour Evolution -- 2.4 Post-processing -- 3 Experimental Results and Analysis -- 3.1 Experimental Dataset -- 3.2 Binarization Results -- 3.3 Performance Evaluation -- 3.4 Comparison with State-of-the-Art Methods -- 3.5 Performance Evaluation Based on OCR -- 4 Conclusion -- References -- Accelerated Stereo Vision Using Nvidia Jetson and Intel AVX -- Abstract -- 1 Introduction -- 1.1 Stereo Depth Estimation -- 1.2 Hardware Selection -- 2 Literature Survey -- 3 Our Implementation -- 3.1 Intel CPU Optimization -- 3.2 Nvidia Jetson Implementation -- 4 Conclusion -- References -- A Novel Machine Annotated Balanced Bangla OCR Corpus -- 1 Introduction -- 2 Literature Review -- 3 Procedure -- 3.1 Data Sources -- 3.2 Layout Analysis -- 3.3 Line Segmentation -- 3.4 Word Segmentation -- 3.5 Character Segmentation -- 3.6 Character Recognition -- 4 Corpus Specification -- 5 Corpus Statistics -- 6 Corpus Balance -- 7 Discussion -- 8 Conclusion -- References -- Generative Adversarial Network for Heritage Image Super Resolution -- 1 Introduction -- 2 SRR via GAN -- 3 Proposed Method -- 3.1 Division of Patches -- 3.2 SRR via GAN Model Using Modified Loss Functions -- 4 Result and Analysis -- 5 Conclusion -- References -- Deep Learning Based Image Enhancement and Black Box Filter Parameter Estimation -- Abstract -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Image Enhancement -- 3.2 Parameter Estimation -- 4 Results -- 4.1 Image Enhancement -- 4.2 Parameter Estimation -- 5 Conclusion and Future Work -- References -- Sign Gesture Recognition from Raw Skeleton Information in 3D Using Deep Learning -- 1 Introduction -- 2 Proposed Methodology -- 2.1 Architecture -- 2.2 BiLSTM -- 2.3 GRU -- 3 Experimental Setup -- 3.1 Dataset Description -- 3.2 Experimental Protocol and Network Hyperparameters.4 Results and Discussion -- 5 Conclusion -- References -- Dual Gradient Feature Pair Based Face Recognition for Aging and Pose Changes -- Abstract -- 1 Introduction -- 2 Proposed DGFP Based Face Recognition -- 2.1 Face Segmentation -- 2.2 DGFP Feature Extraction -- 2.3 Feature Matching and Recognition -- 3 Experimental Results -- 4 Conclusion -- References -- Dynamic User Interface Composition -- Abstract -- 1 Introduction -- 2 Related Work -- 2.1 AI in User Interfaces -- 2.2 Quantifying Aesthetics -- 2.3 Region Identification Using Saliency -- 2.4 Ground Truth Estimation -- 2.5 Agreement Calculation -- 3 Methodology -- 3.1 Dataset Creation -- 3.2 Ground Truth Estimation -- 3.3 Agreement Calculation -- 3.4 Target Accuracy -- 4 Proposed Model -- 4.1 Model Architecture -- 4.2 Loss Function -- 5 Results -- 6 Conclusion -- References -- Lightweight Photo-Realistic Style Transfer for Mobile Devices -- Abstract -- 1 Introduction -- 2 Related Work -- 3 Proposed Method -- 3.1 Network Optimization -- 3.2 Photo-Realistic Smoothing -- 4 Experimental Results -- 4.1 On-Device Implementation -- 5 Conclusion and Future Work -- References -- Cricket Stroke Recognition Using Hard and Soft Assignment Based Bag of Visual Words -- 1 Introduction -- 2 Literature Survey -- 3 Methodology -- 3.1 Feature Extraction -- 3.2 Bag of Visual Words (BoV) -- 4 Experimentation -- 5 Results and Discussion -- 6 Conclusion -- References -- Multi-lingual Indian Text Detector for Mobile Devices -- 1 Introduction -- 2 Related Works -- 3 Proposed Scheme -- 3.1 YOLO V3-Tiny -- 3.2 YOLO V4-Tiny -- 4 Experimental Results -- 5 Conclusion -- References -- Facial Occlusion Detection and Reconstruction Using GAN -- 1 Introduction -- 2 Literature Review -- 2.1 Face De-occlusion -- 2.2 Image Restoration -- 3 Proposed Methodology -- 3.1 Landmark Generation Network -- 3.2 Image Completion Network.3.3 Loss Function -- 4 Experiment and Results -- 4.1 Training Strategy -- 4.2 Experiments -- 4.3 Results -- 5 Conclusion -- References -- Ayurvedic Medicinal Plants Identification: A Comparative Study on Feature Extraction Methods -- Abstract -- 1 Introduction -- 2 Materials and Methods -- 2.1 MepcoTropicLeaf Database -- 2.2 Feature Extraction Methods -- 3 Experiment Results and Discussion -- 4 Conclusion -- Acknowledgement -- References -- Domain Knowledge Embedding Based Multimodal Intent Analysis in Artificial Intelligence Camera -- Abstract -- 1 Introduction -- 2 Prior Work -- 3 Taxonomies -- 4 Dataset -- 5 Proposed Method -- 5.1 Baseline Model -- 5.2 Domain Knowledge Embedding Based Model -- 6 Results -- 7 Conclusion -- References -- Age and Gender Prediction Using Deep CNNs and Transfer Learning -- Abstract -- 1 Introduction -- 2 Related Work -- 3 Methodology -- 3.1 Dataset -- 3.2 Deep CNNs -- 3.3 Transfer Learning -- 4 Evaluation -- 5 Experimentation and Results -- 5.1 Deep CNNs -- 5.2 Transfer Learning -- 6 Conclusion -- References -- Text Line Segmentation: A FCN Based Approach -- 1 Introduction -- 2 Related Work -- 3 Proposed Work -- 3.1 Pre-processing -- 3.2 Preparation of Input to the Netwrok -- 3.3 Network Architecture -- 3.4 Training -- 3.5 Merging the Network Outputs -- 3.6 Post-processing -- 4 Experimental Results -- 5 Conclusion -- References -- Precise Recognition of Vision Based Multi-hand Signs Using Deep Single Stage Convolutional Neural Network -- Abstract -- 1 Introduction -- 2 Literature Review -- 3 Methodology -- 4 Experimental Overview -- 4.1 Model Implementation -- 4.2 Model Training -- 4.3 Model Testing and Prediction -- 5 Results and Discussion -- 6 Conclusion -- Acknowledgement -- References -- Human Gait Abnormality Detection Using Low Cost Sensor Technology -- 1 Introduction -- 2 Related Work.3 Data Collection Procedure.Communications in computer and information science ;1377.Computer visionCongressesImage processingCongressesComputer visionComputer visionImage processingComputer vision.006.37Singh Satish KumarMiAaPQMiAaPQMiAaPQBOOK996464393003316Computer vision and image processing1901686UNISA