Search results for: image dataset

Authors: Keyi Wang

Abstract:

Similar to the touchscreen, hand gesture based human-computer interaction (HCI) is a technology that could allow people to perform a variety of tasks faster and more conveniently. This paper proposes a training method of an image-based hand gesture image and video clip recognition system using a CNN (Convolutional Neural Network) with a dataset. A dataset containing 6 hand gesture images is used to train a 2D CNN model. ~98% accuracy is achieved. Furthermore, a 3D CNN model is trained on a dataset containing 4 hand gesture video clips resulting in ~83% accuracy. It is demonstrated that a Cozmo robot loaded with pre-trained models is able to recognize static and dynamic hand gestures.

Keywords: deep learning, hand gesture recognition, computer vision, image processing

Procedia PDF Downloads 145

Authors: Ali Akyar, Osman Serdar Gedik

Abstract:

Depth estimation is a challenging computer vision task that involves estimating the distance between objects in a scene and the camera. It predicts how far each pixel in the 2D image is from the capturing point. There are some important Monocular Depth Estimation (MDE) studies that are based on Vision Transformers (ViT). We benchmark three major studies. The first work aims to build a simple and powerful foundation model that deals with any images under any condition. The second work proposes a method by mixing multiple datasets during training and a robust training objective. The third work combines generalization performance and state-of-the-art results on specific datasets. Although there are studies with thermal images too, we wanted to benchmark these three non-thermal, state-of-the-art studies with a hybrid image dataset which is taken by Multi-Spectral Dynamic Imaging (MSX) technology. MSX technology produces detailed thermal images by bringing together the thermal and visual spectrums. Using this technology, our dataset images are not blur and poorly detailed as the normal thermal images. On the other hand, they are not taken at the perfect light conditions as RGB images. We compared three methods under test with our thermal dataset which was not done before. Additionally, we propose an image enhancement deep learning model for thermal data. This model helps extract the features required for monocular depth estimation. The experimental results demonstrate that, after using our proposed model, the performance of these three methods under test increased significantly for thermal image depth prediction.

Keywords: monocular depth estimation, thermal dataset, benchmarking, vision transformers

Procedia PDF Downloads 35

Authors: Atuhurra Jesse, N'guessan Yves-Roland Douha, Pabitra Lenka

Abstract:

The detection of diseases within plants has attracted a lot of attention from computer vision enthusiasts. Despite the progress made to detect diseases in many plants, there remains a research gap to train image classifiers to detect the cacao swollen shoot virus disease or CSSVD for short, pertinent to cacao plants. This gap has mainly been due to the unavailability of high quality labeled training data. Moreover, institutions have been hesitant to share their data related to CSSVD. To fill these gaps, image classifiers to detect CSSVD-infected cacao plants are presented in this study. The classifiers are based on VGG16, ResNet50 and Vision Transformer (ViT). The image classifiers are evaluated on a recently released and publicly accessible KaraAgroAI Cocoa dataset. The best performing image classifier, based on ResNet50, achieves 95.39\% precision, 93.75\% recall, 94.34\% F1-score and 94\% accuracy on only 20 epochs. There is a +9.75\% improvement in recall when compared to previous works. These results indicate that the image classifiers learn to identify cacao plants infected with CSSVD.

Keywords: CSSVD, image classification, ResNet50, vision transformer, KaraAgroAI cocoa dataset

Procedia PDF Downloads 107

Authors: Fatma El-Zahraa El-Taher, Ayman Taha, Jane Courtney, Susan Mckeever

Abstract:

Understanding road networks plays an important role in navigation applications such as self-driving vehicles and route planning for individual journeys. Intersections of roads are essential components of road networks. Understanding the features of an intersection, from a simple T-junction to larger multi-road junctions, is critical to decisions such as crossing roads or selecting the safest routes. The identification and profiling of intersections from satellite images is a challenging task. While deep learning approaches offer the state-of-the-art in image classification and detection, the availability of training datasets is a bottleneck in this approach. In this paper, a labelled satellite image dataset for the intersection recognition problem is presented. It consists of 14,692 satellite images of Washington DC, USA. To support other users of the dataset, an automated download and labelling script is provided for dataset replication. The challenges of construction and fine-grained feature labelling of a satellite image dataset is examined, including the issue of how to address features that are spread across multiple images. Finally, the accuracy of the detection of intersections in satellite images is evaluated.

Keywords: satellite images, remote sensing images, data acquisition, autonomous vehicles

Procedia PDF Downloads 149

Authors: Toktam Khatibi

Abstract:

Self-supervised learning (SSL) includes machine learning models that are trained on one aspect and/or one part of the input to learn other aspects and/or part of it. SSL models are divided into two different categories, including pre-text task-based models and contrastive learning ones. Pre-text tasks are some auxiliary tasks learning pseudo-labels, and the trained models are further fine-tuned for downstream tasks. However, one important disadvantage of SSL using pre-text task solving is defining an appropriate pre-text task for each image dataset with a variety of image modalities. Therefore, it is required to design an appropriate pretext task automatically for each dataset and each downstream task. To the best of our knowledge, the automatic designing of pretext tasks for image analysis has not been considered yet. In this paper, we present a framework based on In-context learning that describes each task based on its input and output data using a pre-trained image transformer. Our proposed method combines the input image and its learned description for optimizing the pre-text task design and its hyper-parameters using Meta-learning models. The representations learned from the pre-text tasks are fine-tuned for solving the downstream tasks. We demonstrate that our proposed framework outperforms the compared ones on unseen tasks and image modalities in addition to its superior performance for previously known tasks and datasets.

Keywords: in-context learning (ICL), meta learning, self-supervised learning (SSL), vision-language domain, transformers

Procedia PDF Downloads 83

Authors: Ksheeraj Sai Vepuri, Nada Attar

Abstract:

We as humans use words with accompanying visual and facial cues to communicate effectively. Classifying facial emotion using computer vision methodologies has been an active research area in the computer vision field. In this paper, we propose a simple method for facial expression recognition that enhances accuracy. We tested our method on the FER-2013 dataset that contains static images. Instead of using Histogram equalization to preprocess the dataset, we used Unsharp Mask to emphasize texture and details and sharpened the edges. We also used ImageDataGenerator from Keras library for data augmentation. Then we used Convolutional Neural Networks (CNN) model to classify the images into 7 different facial expressions, yielding an accuracy of 69.46% on the test set. Our results show that using image preprocessing such as the sharpening technique for a CNN model can improve the performance, even when the CNN model is relatively simple.

Keywords: facial expression recognittion, image preprocessing, deep learning, CNN

Procedia PDF Downloads 146

Authors: Tonislav Ivanov, Oleksii Nedashkivskyi, Denis Babeshko, Vadim Pinskiy, Matthew Putman

Abstract:

Training a machine learning model for object detection that generalizes well is known to benefit from a training dataset with diverse examples. However, training datasets usually contain many repeats of common examples of a class and lack rarely seen examples. This is due to the process commonly used during human annotation where a person would proceed sequentially through a list of images labeling a sufficiently high total number of examples. Instead, the method presented involves an active process where, after the initial labeling of several images is completed, the next subset of images for labeling is selected by an algorithm. This process of algorithmic image selection and manual labeling continues in an iterative fashion. The algorithm used for the image selection is a deep learning algorithm, based on the U-shaped architecture, which quantifies the presence of unseen data in each image in order to find images that contain the most novel examples. Moreover, the location of the unseen data in each image is highlighted, aiding the labeler in spotting these examples. Experiments performed using semiconductor wafer data show that labeling a subset of the data, curated by this algorithm, resulted in a model with a better performance than a model produced from sequentially labeling the same amount of data. Also, similar performance is achieved compared to a model trained on exhaustive labeling of the whole dataset. Overall, the proposed approach results in a dataset that has a diverse set of examples per class as well as more balanced classes, which proves beneficial when training a deep learning model.

Keywords: computer vision, deep learning, object detection, semiconductor

Procedia PDF Downloads 139

Authors: Shelby Simpson, William Stanley, Namir Naba, Xiaodi Wang

Abstract:

Both t-SNE and UMAP are brand new state of art tools to predominantly preserve the local structure that is to group neighboring data points together, which indeed provides a very informative visualization of heterogeneity in our data. In this research, we develop a t-SNE and UMAP base neural network image classification algorithm to embed the original dataset to a corresponding low dimensional dataset as a preprocessing step, then use this embedded database as input to our specially designed neural network classifier for image classification. We use the fashion MNIST data set, which is a labeled data set of images of clothing objects in our experiments. t-SNE and UMAP are used for dimensionality reduction of the data set and thus produce low dimensional embeddings. Furthermore, we use the embeddings from t-SNE and UMAP to feed into two neural networks. The accuracy of the models from the two neural networks is then compared to a dense neural network that does not use embedding as an input to show which model can classify the images of clothing objects more accurately.

Keywords: t-SNE, UMAP, fashion MNIST, neural networks

Procedia PDF Downloads 201

Authors: Vahid Anari, Leila Shahmohammadi

Abstract:

Diagnosing and interpreting manually from a large cohort dataset of immunohistochemically stained tissue of tumors using an optical microscope involves subjectivity and also is tedious for pathologist specialists. Moreover, digital pathology today represents more of an evolution than a revolution in pathology. In this paper, we develop and test an unsupervised algorithm that can automatically enhance the IHC image of a meningioma tumor and classify cells into positive (proliferative) and negative (normal) cells. A dataset including 150 images is used to test the scheme. In addition, a new adaptive color image enhancement method is proposed based on a vector directional filter (VDF) and statistical properties of filtering the window. Since the cells are distinguishable by the human eye, the accuracy and stability of the algorithm are quantitatively compared through application to a wide variety of real images.

Keywords: digital pathology, cell segmentation, immunohistochemically, noise reduction

Procedia PDF Downloads 70

Authors: Subodh Chandra Shakya, Badal Shrestha, Suni Thapa, Ashutosh Chauhan, Saugat Adhikari

Abstract:

To tackle our problem of clothes identification, we used different architectures of Convolutional Neural Networks. Among different architectures, the outcome from Inception ResNet V2 and MobileNet V2 seemed promising. On comparison of the metrices, we observed that the Inception ResNet V2 slightly outperforms MobileNet V2 for this purpose. So this paper of ours proposes the cloth identifier using Inception ResNet V2 and also contains the comparison between the outcome of ResNet V2 and MobileNet V2. The document here contains the results and findings of the research that we performed on the DeepFashion Dataset. To improve the dataset, we used different image preprocessing techniques like image shearing, image rotation, and denoising. The whole experiment was conducted with the intention of testing the efficiency of convolutional neural networks on cloth identification so that we could develop a reliable system that is good enough in identifying the clothes worn by the users. The whole system can be integrated with some kind of recommendation system.

Keywords: inception ResNet, convolutional neural net, deep learning, confusion matrix, data augmentation, data preprocessing

Procedia PDF Downloads 189

Authors: Ahmad FAROOQ, Xinfeng Zhang, Fahad Sabah, Raheem Sarwar

Abstract:

In Traditional Chinese Medicine, tongue diagnosis is a vital inspection tool (TCM). In this study, we explore the potential of machine learning in tongue diagnosis. It begins with the cataloguing of the various classifications and characteristics of the human tongue. We infer 24 kinds of tongues from the material and coating of the tongue, and we identify 21 attributes of the tongue. The next step is to apply machine learning methods to the tongue dataset. We use the Weka machine learning platform to conduct the experiment for performance analysis. The 457 instances of the tongue dataset are used to test the performance of five different machine learning methods, including SVM, Random Forests, Decision Trees, and Naive Bayes. Based on accuracy and Area under the ROC Curve, the Support Vector Machine algorithm was shown to be the most effective for tongue diagnosis (AUC).

Keywords: medical imaging, image retrieval, machine learning, tongue

Procedia PDF Downloads 85

Authors: Lan Du, Yan Wang, Hui Dai

Abstract:

Synthetic aperture radar (SAR) image change detection has recently become a challenging problem owing to the existence of speckle noises. In this paper, an unsupervised distribution-free change detection for SAR image based on scale-invariant feature transform (SIFT) keypoints and segmentation is proposed. Firstly, the noise-robust SIFT keypoints which reveal the blob-like structures in an image are extracted in the log-ratio image to reduce the detection range. Then, different from the traditional change detection which directly obtains the change-detection map from the difference image, segmentation is made around the extracted keypoints in the two original multitemporal SAR images to obtain accurate changed region. At last, the change-detection map is generated by comparing the two segmentations. Experimental results on the real SAR image dataset demonstrate the effectiveness of the proposed method.

Keywords: change detection, Synthetic Aperture Radar (SAR), Scale-Invariant Feature Transformation (SIFT), segmentation

Procedia PDF Downloads 390

Authors: Vaishnavi Shivde, Shrishti Sinha, Trapti Mishra

Abstract:

Traffic light detection from a moving vehicle is an important technology both for driver safety assistance functions as well as for autonomous driving in the city. This paper proposed a deep-learning-based traffic light recognition method that consists of a pixel-wise image segmentation technique and a fully convolutional network i.e., UNET architecture. This paper has used a method for detecting the position and recognizing the state of the traffic lights in video sequences is presented and evaluated using Traffic Light Dataset which contains masked traffic light image data. The first stage is the detection, which is accomplished through image processing (image segmentation) techniques such as image cropping, color transformation, segmentation of possible traffic lights. The second stage is the recognition, which means identifying the color of the traffic light or knowing the state of traffic light which is achieved by using a Convolutional Neural Network (UNET architecture).

Keywords: traffic light detection, image segmentation, machine learning, classification, convolutional neural networks

Procedia PDF Downloads 177

Authors: Yalong Jiang, Zheru Chi

Abstract:

In this paper, we study the factors which determine the capacity of a Convolutional Neural Network (CNN) model and propose the ways to evaluate and adjust the capacity of a CNN model for best matching to a specific pattern recognition task. Firstly, a scheme is proposed to adjust the number of independent functional units within a CNN model to make it be better fitted to a task. Secondly, the number of independent functional units in the capsule network is adjusted to fit it to the training dataset. Thirdly, a method based on Bayesian GAN is proposed to enrich the variances in the current dataset to increase its complexity. Experimental results on the PASCAL VOC 2010 Person Part dataset and the MNIST dataset show that, in both conventional CNN models and capsule networks, the number of independent functional units is an important factor that determines the capacity of a network model. By adjusting the number of functional units, the capacity of a model can better match the complexity of a dataset.

Keywords: CNN, convolutional neural network, capsule network, capacity optimization, character recognition, data augmentation, semantic segmentation

Procedia PDF Downloads 156

Authors: Vinay P., Srinivas K. S.

Abstract:

Low Light Image Enhancement is of utmost impor- tance in computer vision based tasks. Applications include vision systems for autonomous driving, night vision devices for defence systems, low light object detection tasks. Many of the existing deep learning methods are resource intensive during the inference step and take considerable time for processing. The algorithm should take considerably less than 41 milliseconds in order to process a real-time video feed with 24 frames per second and should be even less for a video with 30 or 60 frames per second. The paper presents a fast and efficient solution which has two main advantages, it has the potential to be used for a real-time video feed, and it can be used in low compute environments because of the lightweight nature. The proposed solution is a pipeline of three steps, the first one is the use of a simple function to map input RGB values to output RGB values, the second is to balance the colors and the final step is to adjust the contrast of the image. Hence a custom dataset is carefully prepared using images taken in low and bright lighting conditions. The preparation of the dataset, the proposed model, the processing time are discussed in detail and the quality of the enhanced images using different methods is shown.

Keywords: low light image enhancement, real-time video, computer vision, machine learning

Procedia PDF Downloads 210

Authors: Ayush Shrivastava, Arpit Chaudhary, Devang Kulshreshtha, Vibhav Prakash Singh, Rajeev Srivastava

Abstract:

Early diagnosis of breast cancer can improve the survival rate by detecting cancer at an early stage. Breast region segmentation is an essential step in the analysis of digital mammograms. Accurate image segmentation leads to better detection of cancer. It aims at separating out Region of Interest (ROI) from rest of the image. The procedure begins with removal of labels, annotations and tags from the mammographic image using morphological opening method. Pectoral Muscle Sliding Window Algorithm (PMSWA) is used for removal of pectoral muscle from mammograms which is necessary as the intensity values of pectoral muscles are similar to that of ROI which makes it difficult to separate out. After removing the pectoral muscle, Dispersed Region Growing Algorithm (DRGA) is used for segmentation of mammogram which disperses seeds in different regions instead of a single bright region. To demonstrate the validity of our segmentation method, 322 mammographic images from Mammographic Image Analysis Society (MIAS) database are used. The dataset contains medio-lateral oblique (MLO) view of mammograms. Experimental results on MIAS dataset show the effectiveness of our proposed method.

Keywords: CAD, dispersed region growing algorithm (DRGA), image segmentation, mammography, pectoral muscle sliding window algorithm (PMSWA)

Procedia PDF Downloads 315

Authors: Yoganathan Karthik, Guhanathan Poravi

Abstract:

Colorizing near-infrared (NIR) images poses unique challenges due to the absence of color information and the nuances in light absorption. In this paper, we present an approach to NIR image colorization utilizing a synthetic dataset generated from visible light images. Our method addresses two major challenges encountered in NIR image colorization: accurately colorizing objects with color variations and avoiding over/under saturation in dimly lit scenes. To tackle these challenges, we propose a Generative Adversarial Network (GAN)-based framework that learns to map NIR images to their corresponding colorized versions. The synthetic dataset ensures diverse color representations, enabling the model to effectively handle objects with varying hues and shades. Furthermore, the GAN architecture facilitates the generation of realistic colorizations while preserving the integrity of dimly lit scenes, thus mitigating issues related to over/under saturation. Experimental results on benchmark NIR image datasets demonstrate the efficacy of our approach in producing high-quality colorizations with improved color accuracy and naturalness. Quantitative evaluations and comparative studies validate the superiority of our method over existing techniques, showcasing its robustness and generalization capability across diverse NIR image scenarios. Our research not only contributes to advancing NIR image colorization but also underscores the importance of synthetic datasets and GANs in addressing domain-specific challenges in image processing tasks. The proposed framework holds promise for various applications in remote sensing, medical imaging, and surveillance where accurate color representation of NIR imagery is crucial for analysis and interpretation.

Keywords: computer vision, near-infrared images, automatic image colorization, generative adversarial networks, synthetic data

Procedia PDF Downloads 46

Authors: Muhammad Ammar, Talha Ali, Abdul Basit, Bakhtawar Rajput, Zobia Sohail

Abstract:

Music note detection is an area of study for the past few years and has its own influence in music file generation from sheet music. We proposed a method to detect music notes on sheet music using basic thresholding and blob detection. Subsequently, we created a notes dictionary using a semi-supervised learning approach. After notes detection, for each test image, the new symbols are added to the dictionary. This makes the notes detection semi-automatic. The experiments are done on images from a dataset and also on the captured images. The developed approach showed almost 100% accuracy on the dataset images, whereas varying results have been seen on captured images.

Keywords: music note, sheet music, optical music recognition, blob detection, thresholding, dictionary generation

Procedia PDF Downloads 184

Authors: Ashis Pradhan, Mohan P. Pradhan

Abstract:

Pattern Matching is useful for recognizing character in a digital image. OCR is one such technique which reads character from a digital image and recognizes them. Line segmentation is initially used for identifying character in an image and later refined by morphological operations like binarization, erosion, thinning, etc. The work discusses a recognition technique that defines a set of morphological operators based on its orientation in a character. These operators are further categorized into groups having similar shape but different orientation for efficient utilization of memory. Finally the characters are recognized in accordance with the occurrence of frequency in hierarchy of significant pattern of those morphological operators and by comparing them with the existing database of each character.

Keywords: binary image, morphological patterns, frequency count, priority, reduction data set and recognition

Procedia PDF Downloads 417

Authors: N. W. U. D. Chathurani, Shlomo Geva, Vinod Chandran, Proboda Rajapaksha

Abstract:

Selecting an appropriate image representation is the most important factor in implementing an effective Content-Based Image Retrieval (CBIR) system. This paper presents a multi-feature fusion approach for efficient CBIR, based on the distance distribution of features and relative feature weights at the time of query processing. It is a simple yet effective approach, which is free from the effect of features' dimensions, ranges, internal feature normalization and the distance measure. This approach can easily be adopted in any feature combination to improve retrieval quality. The proposed approach is empirically evaluated using two benchmark datasets for image classification (a subset of the Corel dataset and Oliva and Torralba) and compared with existing approaches. The performance of the proposed approach is confirmed with the significantly improved performance in comparison with the independently evaluated baseline of the previously proposed feature fusion approaches.

Keywords: feature fusion, image retrieval, membership function, normalization

Procedia PDF Downloads 348

Authors: Marie Alaghband, Niloofar Yousefi, Ivan Garibay

Abstract:

Facial expressions are important parts of both gesture and sign language recognition systems. Despite the recent advances in both fields, annotated facial expression datasets in the context of sign language are still scarce resources. In this manuscript, we introduce an annotated sequenced facial expression dataset in the context of sign language, comprising over 3000 facial images extracted from the daily news and weather forecast of the public tv-station PHOENIX. Unlike the majority of currently existing facial expression datasets, FePh provides sequenced semi-blurry facial images with different head poses, orientations, and movements. In addition, in the majority of images, identities are mouthing the words, which makes the data more challenging. To annotate this dataset we consider primary, secondary, and tertiary dyads of seven basic emotions of "sad", "surprise", "fear", "angry", "neutral", "disgust", and "happy". We also considered the "None" class if the image’s facial expression could not be described by any of the aforementioned emotions. Although we provide FePh as a facial expression dataset of signers in sign language, it has a wider application in gesture recognition and Human Computer Interaction (HCI) systems.

Keywords: annotated facial expression dataset, gesture recognition, sequenced facial expression dataset, sign language recognition

Procedia PDF Downloads 162

Authors: Yeliz Karaca, Rana Karabudak

Abstract:

Multifractal Denoising techniques are used in the identification of significant attributes by removing the noise of the dataset. Magnetic resonance (MR) image technique is the most sensitive method so as to identify chronic disorders of the nervous system such as Multiple Sclerosis. MRI and Expanded Disability Status Scale (EDSS) data belonging to 120 individuals who have one of the subgroups of MS (Relapsing Remitting MS (RRMS), Secondary Progressive MS (SPMS), Primary Progressive MS (PPMS)) as well as 19 healthy individuals in the control group have been used in this study. The study is comprised of the following stages: (i) L2 Norm Multifractal Denoising technique, one of the multifractal technique, has been used with the application on the MS data (MRI and EDSS). In this way, the new dataset has been obtained. (ii) The new MS dataset obtained from the MS dataset and L2 Multifractal Denoising technique has been applied to the K-Means and Fuzzy C Means clustering algorithms which are among the unsupervised methods. Thus, the clustering performances have been compared. (iii) In the identification of significant attributes in the MS dataset through the Multifractal denoising (L2 Norm) technique using K-Means and FCM algorithms on the MS subgroups and control group of healthy individuals, excellent performance outcome has been yielded. According to the clustering results based on the MS subgroups obtained in the study, successful clustering results have been obtained in the K-Means and FCM algorithms by applying the L2 norm of multifractal denoising technique for the MS dataset. Clustering performance has been more successful with the MS Dataset (L2_Norm MS Data Set) K-Means and FCM in which significant attributes are obtained by applying L2 Norm Denoising technique.

Keywords: clinical decision support, clustering algorithms, multiple sclerosis, multifractal techniques

Procedia PDF Downloads 173

Authors: Hozaifa Zaki, Ghada Soliman

Abstract:

In this paper, we introduced an approach for Automated Dotted Arabic Expiration Date Extraction using Optimized Convolutional Autoencoder (ARABEX) with bidirectional LSTM. This approach is used for translating the Arabic dot-matrix expiration dates into their corresponding filled-in dates. A custom lightweight Convolutional Recurrent Neural Network (CRNN) model is then employed to extract the expiration dates. Due to the lack of available dataset images for the Arabic dot-matrix expiration date, we generated synthetic images by creating an Arabic dot-matrix True Type Font (TTF) matrix to address this limitation. Our model was trained on a realistic synthetic dataset of 3287 images, covering the period from 2019 to 2027, represented in the format of yyyy/mm/dd. We then trained our custom CRNN model using the generated synthetic images to assess the performance of our model (ARABEX) by extracting expiration dates from the translated images. Our proposed approach achieved an accuracy of 99.4% on the test dataset of 658 images, while also achieving a Structural Similarity Index (SSIM) of 0.46 for image translation on our dataset. The ARABEX approach demonstrates its ability to be applied to various downstream learning tasks, including image translation and reconstruction. Moreover, this pipeline (ARABEX+CRNN) can be seamlessly integrated into automated sorting systems to extract expiry dates and sort products accordingly during the manufacturing stage. By eliminating the need for manual entry of expiration dates, which can be time-consuming and inefficient for merchants, our approach offers significant results in terms of efficiency and accuracy for Arabic dot-matrix expiration date recognition.

Keywords: computer vision, deep learning, image processing, character recognition

Procedia PDF Downloads 85

Authors: Xulu Yao, Moi Hoon Yap, Yanlong Zhang

Abstract:

As a branch of artificial neural network, deep learning is widely used in the field of image recognition, but the lack of its dataset leads to imperfect model learning. By analysing the data scale requirements of deep learning and aiming at the application in GUI generation, it is found that the collection of GUI dataset is a time-consuming and labor-consuming project, which is difficult to meet the needs of current deep learning network. To solve this problem, this paper proposes a semi-supervised deep learning model that relies on the original small-scale datasets to produce a large number of reliable data sets. By combining the cyclic neural network with the generated countermeasure network, the cyclic neural network can learn the sequence relationship and characteristics of data, make the generated countermeasure network generate reasonable data, and then expand the Rico dataset. Relying on the network structure, the characteristics of collected data can be well analysed, and a large number of reasonable data can be generated according to these characteristics. After data processing, a reliable dataset for model training can be formed, which alleviates the problem of dataset shortage in deep learning.

Keywords: GUI, deep learning, GAN, data augmentation

Procedia PDF Downloads 187

Authors: Hijratullah Sharifzada, You Wang, Said Ikram Sadat, Hamza Javed, Khalid Akhunzada, Sidra Javed, Sadiq Khan

Abstract:

In the construction industry, safety is of paramount importance given the complex and dynamic nature of construction sites, which are prone to various hazards like falls from heights, being hit by falling objects, and structural collapses. Traditional safety management strategies such as manual inspections and safety training have shown significant limitations. This study presents an intelligent monitoring and analysis system for construction site safety based on an image dataset. A specifically designed Construction Site Safety Image Dataset, comprising 10 distinct classes of objects commonly found on sites, is utilized and divided into training, validation, and test subsets. InceptionV3 and MobileNetV2 are chosen as pre-trained models for feature extraction and are modified through truncation and compression to better suit the task. A novel Feature Fusion architecture is introduced, integrating these modified models along with a Squeeze-and-Excitation block. Experimental results demonstrate that the proposed model achieves a mean Average Precision (mAP) of 0.81 at an IoU threshold of 0.5, with high accuracies for classes like "Safety Cone" (91%) and "Machinery" (93%) but relatively lower accuracy for "Vehicle" (57%). The training process exhibits smooth convergence, and compared to prior methods such as YOLOv4 and SSD, the proposed framework shows superiority in precision and recall. Despite its achievements, the system has limitations, including reliance on visual data and dataset imbalance. Future research directions involve incorporating multi-modal data, conducting real-world deployments, and optimizing for edge deployment, aiming to further enhance construction site safety.

Keywords: construction site safety, intelligent monitoring system, image dataset, InceptionV3, MobileNetV2, feature fusion, squeeze-and-excitation block, mean average precision, object detection

Procedia PDF Downloads 3

Authors: Li Zhou, Yanmei Fang

Abstract:

With current technology, images and videos can be obtained more easily than ever. It is so easy to manipulate these digital multimedia information when obtained, and that the content or source of the image and video could be easily tampered. In this paper, we propose to identify the image and video frame by the texture-based approach, e.g. Markov Transition Probability (MTP), which is in space domain, DCT domain and DWT domain, respectively. In the experiment, image and video frame database is constructed, and is used to train and test the classifier Support Vector Machine (SVM). Experiment results show that the texture-based approach has good performance. In order to verify the experiment result, and testify the universality and robustness of algorithm, we build a random testing dataset, the random testing result is in keeping with above experiment.

Keywords: multimedia forensics, video frame, LBP, MTP, SVM

Procedia PDF Downloads 430

Authors: Shreya Pare, Parvin Akhter

Abstract:

Remotely sensing image classification becomes a very challenging task owing to the high dimensionality of hyperspectral images. The pixel-wise classification methods fail to take the spatial structure information of an image. Therefore, to improve the performance of classification, spatial information can be integrated into the classification process. In this paper, the multilevel thresholding algorithm based on a modified fuzzy entropy function is used to perform the segmentation of hyperspectral images. The fuzzy parameters of the MFE function have been optimized by using a new meta-heuristic algorithm based on the Tree-Search algorithm. The segmented image is classified by a large distribution machine (LDM) classifier. Experimental results are shown on a hyperspectral image dataset. The experimental outputs indicate that the proposed technique (MFE-TSA-LDM) achieves much higher classification accuracy for hyperspectral images when compared to state-of-art classification techniques. The proposed algorithm provides accurate segmentation and classification maps, thus becoming more suitable for image classification with large spatial structures.

Keywords: classification, hyperspectral images, large distribution margin, modified fuzzy entropy function, multilevel thresholding, tree search algorithm, hyperspectral image classification using tree search algorithm

Procedia PDF Downloads 182

Authors: Qasim Munye, Juned Islam, Haseeb Qureshi, Syed Jung

Abstract:

Lung nodules are commonly identified in computed tomography (CT) scans by experienced radiologists at a relatively late stage. Early diagnosis can greatly increase survival. We propose using a pix2pix conditional generative adversarial network to generate realistic images simulating early-stage lung nodule growth. We have applied deep images prior to 2341 slices from 895 computed tomography (CT) scans from the Lung Image Database Consortium (LIDC) dataset to generate pseudo-healthy medical images. From these images, 819 were chosen to train a pix2pix network. We observed that for most of the images, the pix2pix network was able to generate images where the nodule increased in size and intensity across epochs. To evaluate the images, 400 generated images were chosen at random and shown to a medical student beside their corresponding original image. Of these 400 generated images, 384 were defined as satisfactory - meaning they resembled a nodule and were visually similar to the corresponding image. We believe that this generated dataset could be used as training data for neural networks to detect lung nodules at an early stage or to improve the accuracy of such networks. This is particularly significant as datasets containing the growth of early-stage nodules are scarce. This project shows that the combination of deep image prior and generative models could potentially open the door to creating larger datasets than currently possible and has the potential to increase the accuracy of medical classification tasks.

Keywords: medical technology, artificial intelligence, radiology, lung cancer

Procedia PDF Downloads 73

Authors: Filip Rusak, Yulia Arzhaeva, Dadong Wang

Abstract:

Machine learning based Computer-Aided Diagnosis (CAD) is gaining much popularity in medical imaging and diagnostic radiology. However, it requires a large amount of high quality and labeled training image datasets. The training images may come from different sources and be acquired from different radiography machines produced by different manufacturers, digital or digitized copies of film radiographs, with various sizes as well as different pixel intensity distributions. In this paper, a content-aware image augmentation method is presented to deal with these variations. The results of the proposed method have been validated graphically by plotting the removed and added seams of pixels on original images. Two different chest X-ray (CXR) datasets are used in the experiments. The CXRs in the datasets defer in size, some are digital CXRs while the others are digitized from analog CXR films. With the proposed content-aware augmentation method, the Seam Carving algorithm is employed to resize CXRs and the corresponding labels in the form of image masks, followed by histogram matching used to normalize the pixel intensities of digital radiography, based on the pixel intensity values of digitized radiographs. We implemented the algorithms, resized the well-known Montgomery dataset, to the size of the most frequently used Japanese Society of Radiological Technology (JSRT) dataset and normalized our digital CXRs for testing. This work resulted in the unified off-the-shelf CXR dataset composed of radiographs included in both, Montgomery and JSRT datasets. The experimental results show that even though the amount of augmentation is large, our algorithm can preserve the important information in lung fields, local structures, and global visual effect adequately. The proposed method can be used to augment training and testing image data sets so that the trained machine learning model can be used to process CXRs from various sources, and it can be potentially used broadly in any medical imaging applications.

Keywords: computer-aided diagnosis, image augmentation, lung segmentation, medical imaging, seam carving

Procedia PDF Downloads 230

Authors: Jaiden X. Schraut

Abstract:

Medical Imaging Segmentation of Chest X-rays is used for the purpose of identification and differentiation of lung cancer, pneumonia, COVID-19, and similar respiratory diseases. Widespread application of computer-supported perception methods into the diagnostic pipeline has been demonstrated to increase prognostic accuracy and aid doctors in efficiently treating patients. Modern models attempt the task of segmentation and classification separately and improve diagnostic efficiency; however, to further enhance this process, this paper proposes a multi-output network that follows a U-Net architecture for image segmentation output and features an additional CNN module for auxiliary classification output. The proposed model achieves a final Jaccard Index of .9634 for image segmentation and a final accuracy of .9600 for classification on the COVID-19 radiography database.

Keywords: chest X-ray, deep learning, image segmentation, image classification

Procedia PDF Downloads 146