Image Classification Problems Research Articles

Compression Supports Spatial Deep Learning

In the last decades, the domain of spatial computing became more and more data driven, especially when using remote sensing-based images. Furthermore, the satellites provide huge amounts of images, so the number of available datasets is increasing. This leads to the need for large storage requirements and high computational costs when estimating the label scene classification problem using deep learning. This consumes and blocks important hardware recourses, energy, and time. In this article, the use of aggressive compression algorithms will be discussed to cut the wasted transmission and resources for selected land cover classification problems. To compare the different compression methods and the classification performance, the satellite image patches are compressed by two methods. The first method is the image quantization of the data to reduce the bit depth. Second is the lossy and lossless compression of images with the use of image file formats, such as JPEG and TIFF. The performance of the classification is evaluated with the use of convolutional neural networks (CNNs) like VGG16. The experiments indicated that not all remote sensing image classification problems improve their performance when taking the full available information into account. Moreover, compression can set the focus on specific image features, leading to fewer storage needs and a reduction in computing time with comparably small costs in terms of quality and accuracy. All in all, quantization and embedding into file formats do support CNNs to estimate the labels of images, by strengthening the features.

Malaria disease detection system based on convolutional neural network (CNN)

Traditionally to diagnose malaria and compute the parasitemia, a microscope is being used, which involves the use of thick and thin blood smears. However, the efficiency in examining the blood smear depends on the expertise of the laboratory technician in classifying the blood smear using the natural eye, through observation on the microscope as this may produce inaccurate reports due to human errors and offer less classification accuracy for novice laboratory technicians with less experience. In order to carry out a highly, accurate diagnosis of malaria. This paper presents a mobile malaria disease detection system based on Convolutional Neural Network (CNN), a class of Deep Learning (DL) algorithm with end-to-end feature extraction and classification. It is highly scalable and offers superior results in image classification problems, this would be used in training a classification model and deploying the model in a Mobile App. The Mobile App would then be used to diagnose malaria by using the device camera to take photo of patient blood smears for the model to classify, and give result output. Structured System Analysis and Design Methodology (SSADM) was adopted in the design of the research. The Malaria diagnosis system was developed using the JAVA Mobile Edition (ME) programming language and Python programming language to train the model and deployed in the developed software. The developed software meets the objectives of the system.

XAI Grad-CAM 기반 궤양병 감귤 이미지 분류 CNN 모델의 점검

By the rapid development of hardware performance and information processing technology, interest in processing unstructured data and creating value is increasing. Various types of AI architectures are being developed and as the decision-making junction of the model increased exponentially, the performance is being improved. However, complex model structure is a major cause of hindering researchers' ease of interpret results and unlike the speed of development of model performance, the progress is slow on explanatory ability.
 Explainable artificial intelligence (XAI) has emerged to solve this problem and decomposes the model's black box to an understandable level to help improve interpretability and reliability. In this research, we approach the ucler disease citrus image classification problem by using CNN model, and the final model showed approximately 97% accuracy. After that, to improve the reliability of the model and to determine the specific area of the image that played a major role in making the final judgment, Gradient-weighted Class Activation Mapping (Grad-CAM), one of the XAI techniques was applied. As a result of the inspection, it was detected that the shape outside the image wasn't distinguished from the object which was greatly affected. So, the unique shape of a specific object was the main cause of misclassification.

Fine-Grained Image Classification Based on Cross-Attention Network

Due to the high similarity of fine-grained image subclasses, small inter-class changes and large intra-class changes are caused, which leads to the difficulty of fine-grained image classification task. However, existing convolutional neural networks have been unable to effectively solve this problem. Aiming at the above-mentioned fine-grained image classification problem, this paper proposes a multi-scale and multi-level ViT model. First, through data augmentation techniques, the accuracy of fine-grained image classification can be effectively improved. Secondly, the small-scale input and large-scale input of the model make the input image have more feature ex-pressions. The subsequent multi-layeredness effectively utilizes the results of the previous layer of ViT, so that the data of the previous layer can be more effectively used in the next layer of ViT. Finally, cross-attention allows the results of two scale inputs to be fused in a reasonable way. The proposed model is competitive with current mainstream state-of-the-art methods on multiple datasets.

Implementation of CNN and ANN for Fashion-MNIST-Dataset using Different Optimizers

Objectives: The paper presents application of convolution neural network and artificial neural network for image classification problem for clothing dataset along with their performance comparison against different optimizers. The major objective of this paper is to perform image classification on fashion-mnist clothing dataset images. Methods: The methods used here are, the traditional ANN and CNN. Here image classification is performed on Fashion-mnist, clothing dataset using CNN and ANN with different optimizers. The performance of the working of ANN and CNN in classifying images from fashion-mnist dataset is compared against different optimizers namely stochastic gradient Descent, Adagrad, RMS prop and Adam optimizer. Findings: The study found that CNN worked better than ANN yielding training accuracy of 95%, 93% and testing accuracy of 91%, 89% when used with Adam and RmsProp respectively. Novelty: The novelty of this work is to present a comparative study of image classification using CNN, ANN using different optimizers, since not many studies or research articles showed the performance comparison of traditional and convolution neural networks in image classification along with different optimizers. Since the real-world scenarios of today require enormous data to be processed, CNN can fit well to diversify applications since they highly reduce the number of parameters to be trained that speeds up the training process. Moreover, to be specific on image classification problems they require the best and most prominent features to be detected and uncovered; this can be achieved using CNN since it has the concept of convolution using filters at its Core. Hence, CNN is highly recommended for such image classification applications than the traditional artificial-neural-networks because of the aforementioned reasons. Keywords: ANN (Artificial NeuralNetworks); CNN (ConvolutionNeural Networks); Optimization; SGD (Stochastic Gradient Descent); RmsProp (Root mean Square propagation); Adam (Adaptive moment estimation); AdaGrad (Adaptive Gradient)

HRC-mCNNs: A Hybrid Regression and Classification Multibranch CNNs for Automatic Meter Reading With Smart Shell

Nowadays, the meter reading is detached into two parts: 1) image collection and 2) image recognition. Most previous researchers perceived meter reading as an image classification problem and obtained impressive classification performance. However, the numerical is also a critical metric of meter measurement and has not been noticed in previous research. This article redefines the meter reading issue as a hybrid procedure of regression and classification and creates a specific model. The resulting algorithm bespeaks the performance of measurement and recognition. The model consists of a hybrid regression and classification loss function and multibranch convolutional neural networks. We construct two data sets to validate the model: 1) normal data set and 2) carry data set, corresponding to classification and numeric accuracy dividedly. The experiments show that the model establishes new state-of-the-art metrics and achieves 0.5312 mean square error (MSE) on numerical precision and 99.98% accuracy on classification accuracy for 3-min training, by over 70 times on MSE and 0.11% on accuracy than the best performing model proposed by a recent study. Furthermore, a production-ready meter-reading system was deployed in a genuine factory with hybrid regression and classification multibranch convolutional neural networks, smart meter shells, and a set of cloud servers.

Convolutional neural network -based phantom image scoring for mammography quality control

BackgroundVisual evaluation of phantom images is an important, but time-consuming part of mammography quality control (QC). Consistent scoring of phantom images over the device’s lifetime is highly desirable. Recently, convolutional neural networks (CNNs) have been applied to a wide range of image classification problems, performing with a high accuracy. The purpose of this study was to automate mammography QC phantom scoring task by training CNN models to mimic a human reviewer.MethodsEight CNN variations consisting of three to ten convolutional layers were trained for detecting targets (fibres, microcalcifications and masses) in American College of Radiology (ACR) accreditation phantom images and the results were compared with human scoring. Regular and artificially degraded/improved QC phantom images from eight mammography devices were visually evaluated by one reviewer. These images were used in training the CNN models. A separate test set consisted of daily QC images from the eight devices and separately acquired images with varying dose levels. These were scored by four reviewers and considered the ground truth for CNN performance testing.ResultsAlthough hyper-parameter search space was limited, an optimal network depth after which additional layers resulted in decreased accuracy was identified. The highest scoring accuracy (95%) was achieved with the CNN consisting of six convolutional layers. The highest deviation between the CNN and the reviewers was found at lowest dose levels. No significant difference emerged between the visual reviews and CNN results except in case of smallest masses.ConclusionA CNN-based automatic mammography QC phantom scoring system can score phantom images in a good agreement with human reviewers, and can therefore be of benefit in mammography QC.

Benchmarking and scaling of deep learning models for land cover image classification

The availability of the sheer volume of Copernicus Sentinel-2 imagery has created new opportunities for exploiting deep learning methods for land use land cover (LULC) image classification at large scales. However, an extensive set of benchmark experiments is currently lacking, i.e. deep learning models tested on the same dataset, with a common and consistent set of metrics, and in the same hardware. In this work, we use the BigEarthNet Sentinel-2 multispectral dataset to benchmark for the first time different state-of-the-art deep learning models for the multi-label, multi-class LULC image classification problem, contributing with an exhaustive zoo of 62 trained models. Our benchmark includes standard Convolution Neural Network architectures, as well as non-convolutional methods, such as Multi-Layer Perceptrons and Vision Transformers. We put to the test EfficientNets and Wide Residual Networks (WRN) architectures, and leverage classification accuracy, training time and inference rate. Furthermore, we propose to use the EfficientNet framework for the compound scaling of a lightweight WRN, by varying network depth, width, and input data resolution. Enhanced with an Efficient Channel Attention mechanism, our scaled lightweight model emerged as the new state-of-the-art. It achieves 4.5% higher averaged F-Score classification accuracy for all 19 LULC classes compared to a standard ResNet50 baseline model, with an order of magnitude less trainable parameters. We provide access to all trained models, along with our code for distributed training on multiple GPU nodes. This model zoo of pre-trained encoders can be used for transfer learning and rapid prototyping in different remote sensing tasks that use Sentinel-2 data, instead of exploiting backbone models trained with data from a different domain, e.g., from ImageNet. We validate their suitability for transfer learning in different datasets of diverse volumes. Our top-performing WRN achieves state-of-the-art performance (71.1% F-Score) on the SEN12MS dataset while being exposed to only a small fraction of the training dataset.

A Filtering Method for LiDAR Point Cloud Based on Multi-Scale CNN with Attention Mechanism

Point cloud filtering is an important prerequisite for three-dimensional surface modeling with high precision based on LiDAR data. To cope with the issues of low filtering accuracy or excessive model complexity in traditional filtering algorithms, this paper proposes a filtering method for LiDAR point cloud based on a multi-scale convolutional neural network incorporated with the attention mechanism. Firstly, a regular image patch centering on each point is constructed based on the elevation information of point clouds. As thus, the point cloud filtering problem is transformed into the image classification problem. Then, considering the ability of multi-scale convolution to extract features at different scales and the potential of the attention mechanism to capture key information in images, a multi-scale convolutional neural network framework is constructed, and the attention mechanism is incorporated to coordinate multi-scale convolution kernel with channel and spatial attention modules. After this, the feature maps of the LiDAR point clouds can be acquired at different scales. For these feature maps, the weights of each channel layer and different spatial regions can be further tuned adaptively, which makes the network training more targeted, thereby improving the model performance for image classification and eventually separating of ground points and non-ground points preferably. Finally, the proposed method is compared with the cloth simulation filtering method (CSF), deep neural network method (DNN), k-nearest neighbor method (KNN), deep convolutional neural network method (DCNN) and scale-irrelevant and terrain-adaptive method (SITA) for the standard ISPRS dataset of point cloud filtering and the filter dataset of Qinghai. The experimental results show that the proposed method can obtain lower classification errors, which proves the superiority of this method in point cloud filtering.

Frame Regularization of a Convolutional Neural Network in Image-Classification Problems

Frame Regularization of a Convolutional Neural Network in Image-Classification Problems

Ensemble learning for the detection of pli-de-passages in the superior temporal sulcus

The surface of the cerebral cortex is very convoluted, with a large number of folds, the cortical sulci. These folds are extremely variable from one individual to another, and this large variability is a problem for many applications in neuroscience and brain imaging. In particular, sulcal geometry (shape) and sulcal topology (branches, number of pieces) are very variable. “Plis de passages” (PPs) or “annectant gyri” can explain part of the topological variability, namely why sulci have a variable number of pieces across subjects. The concept of PPs was first introduced by Gratiolet (1854) to describe transverse gyri that interconnect both sides of a sulcus, that are frequently buried in the depth of sulci, and that are sometimes apparent on the cortical surface, hence seemingly interrupting the course of sulci and separating them in several pieces. Nevertheless, the difficulty of identifying PPs and the lack of systematic methods to automatically detect them has limited their use. However, based on a recent characterization of PPs in the superior temporal sulcus, we present here a method to automatically detect PPs in the superior temporal sulcus. Local morphology within the sulcus is characterized using cortical surface profiling, and the three-dimensional PP recognition problem is performed as a two-dimensional image classification problem with class-imbalance. This is solved by using an ensemble support vector machine model (EnsSVM) with a rebalancing strategy. Cross validation and quantitative experimental results on an external dataset show the effectiveness and robustness of our approach.

DLA-H: A Deep Learning Accelerator for Histopathologic Image Classification.

It is more than a decade since machine learning and especially its leading subtype deep learning have become one of the most interesting topics in almost all areas of science and industry. In numerous contexts, at least one of the applications of deep learning is utilized or is going to be utilized. Using deep learning for image classification is now very popular and widely used in various use cases. Many types of research in medical sciences have been focused on the advantages of deep learning for image classification problems. Some recent researches show more than 90% accuracy for breast tissue classification which is a breakthrough. A huge number of computations in deep neural networks are considered a big challenge both from software and hardware point of view. From the architectural perspective, this big amount of computing operations will result in high power consumption and computation runtime. This led to the emersion of deep learning accelerators which are designed mainly for improving performance and energy efficiency. Data reuse and localization are two great opportunities for achieving energy-efficient computations with lower runtime. Data flows are mainly designed based on these important parameters. In this paper, DLA-H and BJS, a deep learning accelerator, and its data flow for histopathologic image classification are proposed. The simulation results with the MAESTRO tool showed 756 cycles for total runtime and [Formula: see text] GFLOPS roofline throughput that is an extreme performance improvement in comparison to current general-purpose deep learning accelerators and data flows.

Semi-supervised adversarial discriminative domain adaptation.

Domain adaptation is a potential method to train a powerful deep neural network across various datasets. More precisely, domain adaptation methods train the model on training data and test that model on a completely separate dataset. The adversarial-based adaptation method became popular among other domain adaptation methods. Relying on the idea of GAN, the adversarial-based domain adaptation tries to minimize the distribution between the training and testing dataset based on the adversarial learning process. We observe that the semi-supervised learning approach can combine with the adversarial-based method to solve the domain adaptation problem. In this paper, we propose an improved adversarial domain adaptation method called Semi-Supervised Adversarial Discriminative Domain Adaptation (SADDA), which can outperform other prior domain adaptation methods. We also show that SADDA has a wide range of applications and illustrate the promise of our method for image classification and sentiment classification problems.

Tackling the over-smoothing problem of CNN-based hyperspectral image classification

Convolutional neural networks (CNNs) are very important deep neural networks for analyzing visual imagery. However, most CNN-based methods have the problem of over-smoothing at boundaries, which is unfavorable for hyperspectral image classification. To address this problem, a spectral-spatial multiscale residual network (SSMRN) by fusing two separate deep spectral features and deep spatial features is proposed to significantly reduce over-smoothing and effectively learn the features of objects. In the implementation of the SSMRN, a multiscale residual convolutional neural network is proposed as a spatial feature extractor and a band grouping-based bi-directional gated recurrent unit is utilized as a spectral feature extractor. Considering that the importance of spectral and spatial features may vary depending on the spatial resolution of images, we combine both features with two weighting factors with different initial values that can be adaptively adjusted during the network training. To evaluate the effectiveness of the SSMRN, extensive experiments are conducted on public benchmark data sets. The proposed method can retain the detailed boundary of different objects and yield competitive results compared with several state-of-the-art methods.

Detection of Image Level Forgery with Various Constraints Using DFDC Full and Sample Datasets

The emergence of advanced machine learning or deep learning techniques such as autoencoders and generative adversarial networks, can generate images known as deepfakes, which astonishingly resemble the realistic images. These deepfake images are hard to distinguish from the real images and are being used unethically against famous personalities such as politicians, celebrities, and social workers. Hence, we propose a method to detect these deepfake images using a light weighted convolutional neural network (CNN). Our research is conducted with Deep Fake Detection Challenge (DFDC) full and sample datasets, where we compare the performance of our proposed model with various state-of-the-art pretrained models such as VGG-19, Xception and Inception-ResNet-v2. Furthermore, we perform the experiments with various resolutions maintaining 1:1 and 9:16 aspect ratios, which have not been explored for DFDC datasets by any other groups to date. Thus, the proposed model can flexibly accommodate various resolutions and aspect ratios, without being constrained to a specific resolution or aspect ratio for any type of image classification problem. While most of the reported research is limited to sample or preview DFDC datasets only, we have also attempted the testing on full DFDC datasets and presented the results. Contemplating the fact that the detailed results and resource analysis for various scenarios are provided in this research, the proposed deepfake detection method is anticipated to pave new avenues for deepfake detection research, that engages with DFDC datasets.

Contrastive learning-based image retrieval for automatic recognition of in situ marine plankton images

Abstract Automatic recognition of in situ marine plankton images has long been treated as an image classification problem in machine learning. However, the deep learning-based classifiers are far from robust when used for predicting actual oceanic data that inevitably has distributional and compositional variations from their training sets. This paper proposes a novel image retrieval-based framework for plankton image recognition, within which supervised contrastive learning is used to train a feature extractor for better image representation, and similarity between the input and a gallery of reference images is compared to determine the identity of queries. We construct a dataset of high-quality in situ dark-field images of plankton and suspended particles to train and test the proposed retrieval model. Experimental results show that the image retrieval method has achieved excellent recognition performance similar to the state-of-the-art classification models on a very imbalanced closed-set, and also exhibited better generalizability in dealing with dataset shift and out-of-distribution issues. In addition, the image retrieval method has also demonstrated great architectural flexibility, bringing practical convenience for its adaptation to complex marine application scenarios. This new recognition framework is expected to enable real-time in situ observation of marine plankton in the actual oceanic underwater environment in the near future.

Federated Learning in Massive MIMO 6G Networks: Convergence Analysis and Communication-Efficient Design

In federated learning (FL), model weights must be updated at local users and the base station (BS). These weights are subjected to uplink (UL) and downlink (DL) transmission errors due to the limited reliability of wireless channels. In this paper, we investigate the impact of imperfections in both UL and DL links. First, for a multi-user massive multi-input-multi-output (mMIMO) 6G network, employing zero-forcing (ZF) and minimum mean-squared-error (MMSE) schemes, we analyze the estimation errors of weights for each round. A tighter convergence bound on the modelling error for the communication efficient FL algorithm is derived of the order of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathcal{O}\left(T^{-1}\sigma_{z}^{2}\right)$</tex-math></inline-formula> , where… <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sigma_z^{2}$</tex-math></inline-formula> denotes the variance of overall communication error including the quantization noise. The analysis shows that the reliability of DL links is more critical than that of UL links; and the transmit power can be varied in training process to reduce energy consumption. We also vary the number of local training steps, average codeword length after quantization and scheduling policy to improve the communication efficiency. Simulations with image classification problems on MNIST, EMNIST and FMNIST datasets verify the derived bound and are useful to infer the minimum SNR required for successful convergence of the FL algorithm.

Data augmentation with improved regularisation and sampling for imbalanced blood cell image classification

Due to progression in cell-cycle or duration of storage, classification of morphological changes in human blood cells is important for correct and effective clinical decisions. Automated classification systems help avoid subjective outcomes and are more efficient. Deep learning and more specifically Convolutional Neural Networks have achieved state-of-the-art performance on various biomedical image classification problems. However, real-world data often suffers from the data imbalance problem, owing to which the trained classifier is biased towards the majority classes and does not perform well on the minority classes. This study presents an imbalanced blood cells classification method that utilises Wasserstein divergence GAN, mixup and novel nonlinear mixup for data augmentation to achieve oversampling of the minority classes. We also present a minority class focussed sampling strategy, which allows effective representation of minority class samples produced by all three data augmentation techniques and contributes to the classification performance. The method was evaluated on two publicly available datasets of immortalised human T-lymphocyte cells and Red Blood Cells. Classification performance evaluated using F1-score shows that our proposed approach outperforms existing methods on the same datasets.

Fish image recognition method based on multi-layer feature fusion convolutional network

Solving the problem of fish image classification is important to conserve fish diversity. This conundrum can be addressed by developing a new fish image classification method based on deep learning by training data with complex backgrounds. To this end, this paper proposes a fusion model, referred to as Tripmix-Net. The backbone network of the proposed model primarily consists of multiscale parallel and improved residual networks that are connected in an alternate manner, and network fusion is used to integrate the information that is extracted from shallow and deep layers. Experiments conducted on the 15-category WildFish fish image dataset verified the efficacy of the proposed Tripmix-Net for classifying same-genus fish images with complex backgrounds. The model achieved an accuracy of 95.31%, which is a significant improvement over traditional methods. The proposed approach serves as a new concept for the fine-grained image classification of fish against complex backgrounds.

NEURAL NETWORK BASED REAL TIME PNEUMONIA DETECTION USING TRANSFER LEARNING AND IMAGE AUGMENTATION

This paper aims to identify the detection of pneumonia disease using chest x-ray images by applying deep learning methods. Deep learning methods have tracked down their applications in different areas, and they are in effect broadly utilized in clinical medicines and diagnostics. To analyze viral/bacterial infections such as pneumonia, the assessment of chest X-ray images is frequently used, and the productivity of diagnosis can be altogether improved with the utilization of Computer-Aided Diagnostic (CAD) frameworks. Deep learning method such as Convolutional Neural Network (CNN) architecture is utilized in this paper for the characterization of chest X-ray images to analyze pneumonia. We have used the chest X-ray image dataset from Kaggle consisting of 4110 images. Image augmentations were performed on the dataset to oversample the dataset for the model to perform better. Then, at that point, we have built a custom CNN model and, also, we have utilized the transfer learning mechanism with CNN by using MobileNetV2 as the base model for the image classification problems. The average classification accuracy for our proposed CNN and MobileNetV2 based transfer learning method was 97%, and 97% for unbalanced and 97%, and 97%, for balanced datasets respectively. The satisfactory outcome of both models can significantly improve the accuracy and speed of pneumonia diagnosis. This would be very helpful in this pandemic situation in developing countries with limited resources and capabilities in the healthcare sector.