Image Classification System Research Articles

Adapted Jacobi Orthogonal Invariant Moments for Image Representation and Recognition

Images recognition and classification require an extraction technique of feature vectors of these images. These vectors must be invariant to the three geometric transformations: rotation, translation and scaling. Several authors used the theory of orthogonal moments to extract the feature vectors of images. Jacobi moments are orthogonal moments, which have been widely applied in imaging and pattern recognition. However, the invariance to rotation of Cartesian Jacobi moments is very difficult to obtain. In this paper, we obtain at first a set of transformed orthogonal Jacobi polynomials, called “Adapted Jacobi polynomials”. Based on these polynomials, a set of orthogonal moments is presented, named adapted Jacobi moments (AJMs). These moments are orthogonal on the rectangle $$\left[ {0, N\left] { \times } \right[0, M} \right],$$ where $$N \times M$$ is the size of the described image. We also provide a new series of feature vectors of images based on adapted Jacobi orthogonal invariants moments, which are a linear combination of geometric moment invariants, where the latest ones are invariant under rotation, translation and scaling of the described image. Based on k-NN algorithm, we apply a new 2D image classification system. We introduce a set of experimental tests in pattern recognition. The obtained results express the efficiency of our method. The performance of these feature vectors is compared with someones extracted from Hu, Legendre and Tchebichef invariant moments using three different 2D image databases: MPEG7-CE shape database, Columbia Object Image Library (COIL-20) database and ORL database. The results of the comparative study show the performance and superiority of our orthogonal invariant moments.

Image classification with deep learning in the presence of noisy labels: A survey

Image classification systems recently made a giant leap with the advancement of deep neural networks. However, these systems require an excessive amount of labeled data to be adequately trained. Gathering a correctly annotated dataset is not always feasible due to several factors, such as the expensiveness of the labeling process or difficulty of correctly classifying data, even for the experts. Because of these practical challenges, label noise is a common problem in real-world datasets, and numerous methods to train deep neural networks with label noise are proposed in the literature. Although deep neural networks are known to be relatively robust to label noise, their tendency to overfit data makes them vulnerable to memorizing even random noise. Therefore, it is crucial to consider the existence of label noise and develop counter algorithms to fade away its adverse effects to train deep neural networks efficiently. Even though an extensive survey of machine learning techniques under label noise exists, the literature lacks a comprehensive survey of methodologies centered explicitly around deep learning in the presence of noisy labels. This paper aims to present these algorithms while categorizing them into one of the two subgroups: noise model based and noise model free methods. Algorithms in the first group aim to estimate the noise structure and use this information to avoid the adverse effects of noisy labels. Differently, methods in the second group try to come up with inherently noise robust algorithms by using approaches like robust losses, regularizers or other learning paradigms.

Ensemble learning of diffractive optical networks

A plethora of research advances have emerged in the fields of optics and photonics that benefit from harnessing the power of machine learning. Specifically, there has been a revival of interest in optical computing hardware due to its potential advantages for machine learning tasks in terms of parallelization, power efficiency and computation speed. Diffractive deep neural networks (D2NNs) form such an optical computing framework that benefits from deep learning-based design of successive diffractive layers to all-optically process information as the input light diffracts through these passive layers. D2NNs have demonstrated success in various tasks, including object classification, the spectral encoding of information, optical pulse shaping and imaging. Here, we substantially improve the inference performance of diffractive optical networks using feature engineering and ensemble learning. After independently training 1252 D2NNs that were diversely engineered with a variety of passive input filters, we applied a pruning algorithm to select an optimized ensemble of D2NNs that collectively improved the image classification accuracy. Through this pruning, we numerically demonstrated that ensembles of N = 14 and N = 30 D2NNs achieve blind testing accuracies of 61.14 ± 0.23% and 62.13 ± 0.05%, respectively, on the classification of CIFAR-10 test images, providing an inference improvement of >16% compared to the average performance of the individual D2NNs within each ensemble. These results constitute the highest inference accuracies achieved to date by any diffractive optical neural network design on the same dataset and might provide a significant leap to extend the application space of diffractive optical image classification and machine vision systems.

Prioritized multi-criteria federated learning

In Machine Learning scenarios, privacy is a crucial concern when models have to be trained with private data coming from users of a service, such as a recommender system, a location-based mobile service, a mobile phone text messaging service providing next word prediction, or a face image classification system. The main issue is that, often, data are collected, transferred, and processed by third parties. These transactions violate new regulations, such as GDPR. Furthermore, users usually are not willing to share private data such as their visited locations, the text messages they wrote, or the photo they took with a third party. On the other hand, users appreciate services that work based on their behaviors and preferences. In order to address these issues, Federated Learning (FL) has been recently proposed as a means to build ML models based on private datasets distributed over a large number of clients, while preventing data leakage. A federation of users is asked to train a same global model on their private data, while a central coordinating server receives locally computed updates by clients and aggregate them to obtain a better global model, without the need to use clients’ actual data. In this work, we extend the FL approach by pushing forward the state-of-the-art approaches in the aggregation step of FL, which we deem crucial for building a high-quality global model. Specifically, we propose an approach that takes into account a suite of client-specific criteria that constitute the basis for assigning a score to each client based on a priority of criteria defined by the service provider. Extensive experiments on two publicly available datasets indicate the merits of the proposed approach compared to standard FL baseline.

Cervical spine trauma: impact of different imaging classification systems in the clinical decision-making.

Background and aim:Considering the high rate of mortality and permanent disability related to vertebral traumas, an early and detailed diagnosis of the trauma and subsequently an immediate and effective intervention are crucial. Cervical vertebral injury classifications guide treatment choice through a severity grade based on radiological information. The purpose of the present study was to define which imaging classification system could provide the best morphological and clinical-surgical correlations for cervical spine traumas.Methods:We retrospectively analyzed patients evaluated for cervical spine trauma at our Institution in the period 2015-2020. Information regarding the morphological examination (using CT and MRI), the neurological evaluation, and the therapeutic management were collected. C3-C7 fractures were classified according to the SLIC and AOSpine criteria; axial lesions were classified according to the modified AOSpine for the C1-C2 compartment and through the Roy-Camille and the Anderson D’Alonzo system for the odontoid process of the axis.Results:29 patients were included in the final study population. Nine patients with axial spine trauma and 21 with subaxial cervical spine trauma. A conservative approach was applied in 16 patients while nine patients underwent neurosurgery. Considering the therapeutical indications provided by the SLIC system, a 76.9% accordance was found for patients with a <4 score, while a 100% concordance was calculated for patients with a >4 score undergoing neurosurgery. Regarding the AOSspine classification, a 28.6% concordance was observed for patients classified group B being treated with a posterior neurosurgical approach, while for patients belonging to subgroup C, considered for anterior neurosurgical approach, a 66.7% accordance was calculated.Conclusions:The study demonstrated a better morphological correlation for the AOSpine classification in subaxial trauma and the AOSpine and Anderson D’Alonzo in axial trauma. The therapeutic indication found a better correlation in the SLIC classification for subaxial trauma and the Anderson D’Alonzo for axial ones. (www.actabiomedica.it)

MR-brain image classification system based on SWT-LBP and ensemble of SVMs

In this paper, we present an efficient magnetic resonance (MR) image classification system. At the first stage, the brain image is decomposed into several subbands using stationary wavelet transform (SWT). Then, local binary patterns (LBP) with reduced histograms are computed for each subband to form several primary feature vectors. Principal components analysis (PCA) followed by linear discriminant analysis (LDA) are then applied to each primary feature vector in order to transform them into new lower-dimension feature vectors. The third stage consists of using an ensemble of support vector machines (SVMs) in order to build voters and make the final decision on the requested image. The designed system is evaluated on 255 brain images with five-fold cross-validation. Experimental results show that the proposed system achieves a classification rate of 99.78% which outperforms the existing brain classification approaches.

Development and Usability Testing of a Consultation System for Diabetic Retinopathy Screening

This study aims to develop a novel web-based decision support system for diabetic retinopathy screening and classification of eye fundus images for medical officers. The research delivers diabetic retinopathy information with a web-based environment according to the needs of the users. The proposed research also intends to evaluate the developed system usability to the target users. The complex characteristics of diabetic retinopathy signs contribute to the difficulty in detecting diabetic retinopathy. Therefore, professional and skilled retinal screeners are required to produce accurate diabetic retinopathy detection and diagnosis. The proposed system assists the communication and consultation among the medical experts in the hospital and the primary health cares located at the health clinics. The agile software development model is the methodology used for the development of this research project. The project collaborates with the Department of Ophthalmology, Hospital Melaka, Malaysia for the medical content expertise and testing. Representative medical officers from Hospital Melaka and all the public health clinics in Melaka were involved in the preliminary study and system testing. This research study consists of a web development producing an interactive web-based application of diabetic retinopathy consultation which comprises image processing and editing features as a core of the system. It is envisaged that this research project will contribute to the management of diabetic retinopathy screening among medical officers.

Optimal Flame Detection of Fires in Videos Based on Deep Learning and the Use of Various Optimizers

Deep learning has recently attracted a lot of attention with the aim of developing a fast, automatic and accurate system for image identification and classification. In this work, the focus was on transfer learning and evaluation of state-of-the-art VGG16 and 19 deep convolutional neural networks for fire image classification from fire images. In this study, five different approaches (Adagrad, Adam, AdaMax, Nadam and Rmsprop) based on the gradient descent methods used in parameter updating were studied. By selecting specific learning rates, training image base proportions, number of recursion (epochs), the advantages and disadvantages of each approach are compared with each other in order to achieve the minimum cost function. The results of the comparison are presented in the tables. In our experiment, Adam optimizers with the VGG16 architecture with 300 and 500 epochs tend to steadily improve their accuracy with increasing number of epochs without deteriorating performance. The optimizers were evaluated on the basis of their AUC of the ROC curve. It achieves a test accuracy of 96%, which puts it ahead of other architectures.

On the Initial Behavior Monitoring Issues in Federated Learning

In Federated Learning (FL), a group of workers participate to build a global model under the coordination of one node, the chief. Regarding the cybersecurity of FL, some attacks aim at injecting the fabricated local model updates into the system. Some defenses are based on malicious worker detection and behavioral pattern analysis. In this context, without timely and dynamic monitoring methods, the chief cannot detect and remove the malicious or unreliable workers from the system. Our work emphasize the urgency to prepare the federated learning process for monitoring and eventually behavioral pattern analysis. We study the information inside the learning process in the early stages of training, propose a monitoring process and evaluate the monitoring period required. The aim is to analyse at what time is it appropriate to start the detection algorithm in order to remove the malicious or unreliable workers from the system and optimise the defense mechanism deployment. We tested our strategy on a behavioral pattern analysis defense applied to the FL process of different benchmark systems for text and image classification. Our results show that the monitoring process lowers false positives and false negatives and consequently increases system efficiency by enabling the distributed learning system to achieve better performance in the early stage of training.

A Novel Method for Quantitative Analysis of C-Reactive Protein Lateral Flow Immunoassays Images via CMOS Sensor and Recurrent Neural Networks.

Objective: To design and implement an easy-to-use, Point-of-Care (PoC) lateral flow immunoassays (LFA) reader and data analysis system, which provides a more in-depth quantitative analysis for LFA images than conventional approaches thereby supporting efficient decision making for potential early risk assessment of cardiovascular disease (CVD). Methods and procedures: A novel end-to-end system was developed including a portable device with CMOS camera integrated with optimized illumination and optics to capture the LFA images produced using high-sensitivity C-Reactive Protein (hsCRP) (concentration level < 5 mg/L). The images were transmitted via WiFi to a back-end server system for image analysis and classification. Unlike common image classification approaches which are based on averaging image intensity from a region-of-interest (ROI), a novel approach was developed which considered the signal along the sample’s flow direction as a time series and, consequently, no need for ROI detection. Long Short-Term Memory (LSTM) networks were deployed for multilevel classification. The features based on Dynamic Time Warping (DTW) and histogram bin counts (HBC) were explored for classification. Results: For the classification of hsCRP, the LSTM outperformed the traditional machine learning classifiers with or without DTW and HBC features performed the best (with mean accuracy of 94%) compared to other features. Application of the proposed method to human plasma also suggests that HBC features from LFA time series performed better than the mean from ROI and raw LFA data. Conclusion: As a proof of concept, the results demonstrate the capability of the proposed framework for quantitative analysis of LFA images and suggest the potential for early risk assessment of CVD. Clinical impact: The hsCRP levels < 5 mg/L were aligned with clinically actionable categories for early risk assessment of CVD. The outcomes demonstrated the real-world applicability of the proposed system for quantitative analysis of LFA images, which is potentially useful for more LFA applications beyond presented in this study.

MR-brain image classification system based on SWT-LBP and ensemble of SVMs

MR-brain image classification system based on SWT-LBP and ensemble of SVMs

Development of a digital image classification system to support technical assistance for Black Sigatoka detection

Abstract A large percentage of Colombia’s economic activity corresponds to the agricultural sector. In this sector, plantains rank second in production and planted area. However this crop is affected by different diseases, among which The Black Sigatoka stands out, caused by the fungus Mycosphaerella fijiensis. The disease highly reduces the production level of the crop and although there are prevention measures that allow reducing the incidence of the disease, there’s a lack of support for small producers in Colombia, who do not have technological tools to support the disease detection processes. This article outlines the development of a support system for the detection of black sigatoka using digital images. For this, a characterization process of the agricultural user is carried out, then, a machine learning methodology is implemented to classify the disease on a mobile device. The support system is validated through laboratory tests, field tests and the feedback from the agricultural user.

Deep Learning Based Mineral Image Classification Combined With Visual Attention Mechanism

Mineral image classification technology based on machine vision is an efficient system for ore sorting. With the development of artificial intelligence and computer technology, the deep learning-based mineral image classification system is gradually applied to ore sorting. However, there is a bottleneck in improving classification accuracy, and the feature extraction ability of the CNNs model is relatively limited for multi-category mineral image classification tasks. Therefore, four visual attention blocks are designed and embedded in the existing CNNs model, and new mineral image classification models based on the visual attention mechanism and CNNs are proposed. Then, referring to the building strategies of the different depth ResNet, we build various CNNs model embedding with attention blocks for mineral image classification and visualize the models by Grad-CAM to observe the change in classification weight distributions and classification weight values. Finally, by using the confusion matrices, this experiment systematically evaluates the classification performance of the proposed models and analyzes the misjudgment rate.

Galactography Combined with Sonogalactography for Improving the Evaluation of Pathological Nipple Discharge

Diagnosing patients with pathological nipple discharge (PND) is controversial, and therefore a standardized diagnosis algorithm is needed. The objective of this study was to investigate the usefulness of galactography (GL) combined with sonogalactography (SGL) for the evaluation of PND patients. A retrospective study was conducted of 51 patients with PND who were evaluated with GL and SGL. The findings from the galactograms of the patients in this study were assigned to different categories of the Galactogram Image Classification System. Additionally, the sensitivity, specificity, and the positive predictive values and negative predictive values of the GL and SGL tests were calculated, considering the gold standard of pathology diagnosis. The results obtained show that GL combined with SGL improved the diagnostic efficiency of ductal lesions, especially for borderline and malignant lesions. Papilloma was diagnosed in 19 cases, and ductal carcinoma in situ in 8 patients. Conclusions: To the best of our knowledge, this is the first study in which the combination of GL and SGL improves the diagnostic efficiency of ductal lesions of patients with PND. A diagnosis algorithm is recommended for women with PND.

Knowledge-based generative adversarial networks for scene understanding in Cultural Heritage

This study shows the results of an interdisciplinary research aimed at devising artificial intelligence methods to support data understanding in Cultural Heritage. The objective is to create automated methods for automatically classifying and successively retrieving photographic data leveraging the current breed of AI methods based on Deep Learning paradigm. In this work, the lack of images to be used for the training of the AI system is addressed by testing an approach based on Generative Adversarial Networks (GANs) for automatically synthesizing unrealistic photos, which, in turn, can be used for training image classification and retrieval systems. Specifically, we propose a method to drive the generation of realistic classical order images using GAN approaches anchored to semantic ontology domain representation. More specifically, the proposed AI leverages the advantages of knowledge-based and data-driven methods and foresees a level image generation: the first one that synthesizes isolated objects corresponding to each ontology concept and the second one that, instead, combines the generated objects according to the spatial information provided by the ontology in order to generate realistic scenes. The resulting images can then be employed to train automated classifiers and more reliable retrieval methods by minimizing the efforts usually required by human operators to manually annotate data. In addition, the proposed generative strategy allows for a deeper understanding of cultural heritage visual data enabling the possibility to enhance design capabilities in simulation contexts.

A System for Brain Image Segmentation and Classification Based on Three-Dimensional Convolutional Neural Network

We consider the problem of fully automatic brain tumor segmentation in MR images containing glioblastomas. We propose a three-Dimensional Convolutional Neural Network (3D-CNN) approach that achieves high performance while being extremely efficient, a balance that existing methods have struggled to achieve. Our 3D-Brain CNN is formed directly on raw image modalities and thus learn a characteristic representation directly from the data. We propose a new cascading architecture with two pathways that each model normal details in tumors. Fully exploiting the convolutional nature of our model also allows us to segment a complete cerebral image in one minute. In experiments on the 2013 and 2015 BRATS challenge dataset; we exhibit that our approach is among the most powerful methods in the literature, while also being very effective.

Relationship between Machine-Learning Image Classification of T2-Weighted Intramedullary Hypointensity on 3 Tesla Magnetic Resonance Imaging and Clinical Outcome in Dogs with Severe Spinal Cord Injury.

Early prognostic information in cases of severe spinal cord injury can aid treatment planning and stratification for clinical trials. Analysis of intraparenchymal signal change on magnetic resonance imaging has been suggested to inform outcome prediction in traumatic spinal cord injury. We hypothesized that intraparenchymal T2-weighted hypointensity would be associated with a lower potential for functional recovery and a higher risk of progressive neurological deterioration in dogs with acute, severe, naturally occurring spinal cord injury. Our objectives were to: 1) demonstrate capacity for machine-learning criteria to identify clinically relevant regions of hypointensity and 2) compare clinical outcomes for cases with and without such regions. A total of 95 dogs with complete spinal cord injury were evaluated. An image classification system, based on Speeded-Up Robust Features (SURF), was trained to recognize individual axial T2-weighted slices that contained hypointensity. The presence of such slices in a given transverse series was correlated with a lower chance of functional recovery (odds ratio [OR], 0.08; confidence interval [CI], 0.02-0.38; p < 10-3) and with a higher risk of neurological deterioration (OR, 0.14; 95% CI, 0.05-0.42; p < 10-3). Identification of intraparenchymal T2-weighted hypointensity in severe, naturally occurring spinal cord injury may be assisted by an image classification tool and is correlated with functional recovery.

End to end system for hazy image classification and reconstruction based on mean channel prior using deep learning network

Outdoor images are having several applications including autonomous vehicles, geo-mapping, and surveillance. It is a common phenomenon that the images captured outdoor are prone to noise, which arises due to natural and manmade extreme atmospheric conditions such as haze, fog, and smog. Importantly in autonomous vehicle navigation, it is very important to recover the ground truth image to get the better decision by the system. Estimation of the transmission map and air-light is very crucial in recovering the ground truth image. In this study, the authors proposed a new method to estimate the transmission map based on a mean channel prior (MCP), which represents the depth map to estimate the transmission map. The authors proposed a deep neural network to identify the hazy image for the further dehazing process. In this study, the authors presented, two novel contributions, first an MCP-based image dehazing and second, a deep neural network-based identification of hazy images as a pre-processing block in the proposed end to end system. The proposed deep learning network using the TensorFlow platform provided validation accuracy of 93.4% for hazy image classification. Finally, the proposed MCP-based dehazing network showed better performance in terms of peak-signal-to-noise ratio, structural similarity index, and computational time than that of existing methods.

Personalized Image Classification by Semantic Embedding and Active Learning.

Currently, deep learning has shown state-of-the-art performance in image classification with pre-defined taxonomy. However, in a more real-world scenario, different users usually have different classification intents given an image collection. To satisfactorily personalize the requirement, we propose an interactive image classification system with an offline representation learning stage and an online classification stage. During the offline stage, we learn a deep model to extract the feature with higher flexibility and scalability for different users’ preferences. Instead of training the model only with the inter-class discrimination, we also encode the similarity between the semantic-embedding vectors of the category labels into the model. This makes the extracted feature adapt to multiple taxonomies with different granularities. During the online session, an annotation task iteratively alternates with a high-throughput verification task. When performing the verification task, the users are only required to indicate the incorrect prediction without giving the exact category label. For each iteration, our system chooses the images to be annotated or verified based on interactive efficiency optimization. To provide a high interactive rate, a unified active learning algorithm is used to search the optimal annotation and verification set by minimizing the expected time cost. After interactive annotation and verification, the new classified images are used to train a customized classifier online, which reflects the user-adaptive intent of categorization. The learned classifier is then used for subsequent annotation and verification tasks. Experimental results under several public image datasets show that our method outperforms existing methods.

New set of adapted Gegenbauer–Chebyshev invariant moments for image recognition and classification

Orthogonal moments and their invariants to geometric transformations for gray-scale images are widely used in many pattern recognition and image processing applications. In this paper, we propose a new set of orthogonal polynomials called adapted Gegenbauer–Chebyshev polynomials (AGC). This new set is used as a basic function to define the orthogonal adapted Gegenbauer–Chebyshev moments (AGCMs). The rotation, scaling, and translation invariant property of (AGCMs) is derived and analyzed. We provide a novel series of feature vectors of images based on the adapted Gegenbauer–Chebyshev orthogonal moments invariants (AGCMIs). We practice a novel image classification system using the proposed feature vectors and the fuzzy k-means classifier. A series of experiments is performed to validate this new set of orthogonal moments and compare its performance with the existing orthogonal moments as Legendre invariants moments, the Gegenbauer and Tchebichef invariant moments using three different image databases: the MPEG7-CE Shape database, the Columbia Object Image Library (COIL-20) database and the ORL-faces database. The obtained results ensure the superiority of the proposed AGCMs over all existing moments in representation and recognition of the images.