Image Block Classification Research Articles

Research on automated rapid diagnosis of distal radius fractures using X-ray images

The primary diagnostic method for distal radius fractures remains conventional X-ray imaging. In emergency situations, rapid assessment and decision-making are often required from non-orthopedic physicians or junior radiologists in the emergency room. Therefore, an accurate and efficient auxiliary diagnostic technology for distal radius fractures is of great significance. This paper combines deep learning with image analysis techniques to propose an effective classification method for types of distal radius fractures. Firstly, an extended U-Net three-level cascade segmentation network is used for precise segmentation of the most critical joint surface and non-joint surface areas for fracture identification. The edge is further optimized using the central pixel image block classification method. Then, fracture identification is performed separately on the joint surface and non-joint surface areas. Finally, based on the classification results of the two types of images, a comprehensive judgment is made to determine normal or ABC fracture types. The experimental X-ray data comes from three tertiary hospitals, with a total of 12,000 images in the training set, 3,000 of each type, and 1,200 images in the test set, 300 of each type. To further demonstrate the generalization performance of the classification experiment, an additional 500 images were added for fracture classification testing, coming from collaborative medical institutions in the United States and Germany. All images were annotated by orthopedic medical experts with more than ten years of experience. The accuracy rates for normal, type A, type B, and type C fractures in the test set were 0.99, 0.92, 0.91, and 0.82, respectively. For orthopedic medical experts, the average recognition accuracy rates were 0.98, 0.90, 0.87, and 0.81. The proposed automatic recognition method overall performs better than experts and can be used for preliminary auxiliary diagnosis of distal radius fractures without expert participation.

Research on automated rapid diagnosis of distal radius fractures using X-ray images

The primary diagnostic method for distal radius fractures remains conventional X-ray imaging. In emergency situations, rapid assessment and decision-making are often required from non-orthopedic physicians or junior radiologists in the emergency room. Therefore, an accurate and efficient auxiliary diagnostic technology for distal radius fractures is of great significance. This paper combines deep learning with image analysis techniques to propose an effective classification method for types of distal radius fractures. Firstly, an extended U-Net three-level cascade segmentation network is used for precise segmentation of the most critical joint surface and non-joint surface areas for fracture identification. The edge is further optimized using the central pixel image block classification method. Then, fracture identification is performed separately on the joint surface and non-joint surface areas. Finally, based on the classification results of the two types of images, a comprehensive judgment is made to determine normal or ABC fracture types. The experimental X-ray data comes from three tertiary hospitals, with a total of 12,000 images in the training set, 3,000 of each type, and 1,200 images in the test set, 300 of each type. To further demonstrate the generalization performance of the classification experiment, an additional 500 images were added for fracture classification testing, coming from collaborative medical institutions in the United States and Germany. All images were annotated by orthopedic medical experts with more than ten years of experience. The accuracy rates for normal, type A, type B, and type C fractures in the test set were 0.99, 0.92, 0.91, and 0.82, respectively. For orthopedic medical experts, the average recognition accuracy rates were 0.98, 0.90, 0.87, and 0.81. The proposed automatic recognition method overall performs better than experts and can be used for preliminary auxiliary diagnosis of distal radius fractures without expert participation.

A Method for Identifying Sports Behaviors in Sports Adversarial Project Training Based on Image Block Classification under the Internet of Things

Abstract There is a need to devise intelligent methods for motion-based image classification to produce accurate results for correct judgment. Existing methods cannot use the image block classification method to recognize the motion behavior of antagonistic sports training, and it is the need of the hour to devise newer methods. The recognition result is not ideal when using the traditional methods. This paper proposes an image block classification method for antagonistic sports training behavior recognition. It classifies the noise sources, quantifies the influence degree of each noise source effectively, and adjusts the noise factor in the denoising method for different noise sources and the specific weight of different noise sources. The adaptive denoising of different noise types of antagonistic sports training image sequences is realized. By using the combination of key frame template selection method and image segmentation method for behavior representation, the features are extracted well. The feature extraction is done on an image region according to the proportion of the number of foreground pixels in each block of the template to the number of pixels in the block. The feature vector is used to form the template. Aiming at the behavior representation and feature extraction based on the features of skeleton joint points, image block classification is used to extract the coordinates of skeleton joint points during human movement. K-means clustering is used to transform them into symbol sequences to represent behavior features that are used for antagonistic sports training behavior recognition. Simulation results show that the proposed method can obtain ideal recognition results and outperforms the existing methods.

Comparative study of deep learning models on the images of biopsy specimens for diagnosis of lung cancer treatment

ObjectiveTo compare the application value of different traditional deep learning models in diagnosing and classifying lung cancer. MethodsAccording to the biopsy samples of our hospital from January 2018 to November 2022, 37 patients treated in this department were selected as the study subjects. Nonsmall and small cell lung cancer biopsy specimens were obtained and stained. Two experienced pathologists diagnosed the biopsy specimens. Multiple in-depth learning methods were used to distinguish between cancer and noncancer biopsies. In this study, we compared the application value of traditional deep learning models in lung cancer diagnosis and Classification. ResultsThe study tested several popular CNN architectures based on image block classification: AlexNet, VGG, ResNet, and SqueezeNet, comparing two types of training schemes: training from scratch and fine-tuning the entire pretrained network. The AUC of the deep learning model is more reasonable (0. 8808–0. 9121). Except for Resnet-50, the AUC of the training is higher than that of the fine-tuning of the whole network. ConclusionDeep learning analysis can accelerate the detection speed of whole section images (WSI) and maintain a similar detection rate with pathologists.

Low-dose CT iterative reconstruction based on image block classification and dictionary learning

Low-dose CT iterative reconstruction based on image block classification and dictionary learning

Human Motion Target Posture Detection Algorithm Using Semi-Supervised Learning in Internet of Things

To address the problem that the traditional human motion attitude detection process is easy to ignore the data calibration, which leads to the problems of long running time, low accuracy and poor detection effect, a human motion target attitude detection algorithm based on semi-supervised learning in the Internet of things environment is proposed. Firstly, human motion target images are collected using the Internet of things (IoT), human motion attitude features are extracted based on the eight-star model, and multi-features are fused to form image blocks of 17-dimensional feature vectors. Then, random fern classifiers are optimized and semi-supervised learning is used to calculate a large number of uncalibrated data in time domain, spatial domain and data. The classifier is trained to complete image block classification. Finally, the classifier parameters are updated iteratively to complete the attitude detection of human motion target. The results show that the proposed algorithm has high accuracy in human motion attitude extraction and multi-feature fusion, and has a high correct classification rate for different feature poses, as high as 92.5%. The average F value of human motion attitude detection is 0.95, the overlap ratio is high and the time is short. The overall performance is good.

Image Restoration Based on Gradual Reweighted Regularization and Low Rank prior

Digital restoration of image with missing data is a basic need for visual communication and industrial applications. In this paper, making full use of priors of low rank and nonlocal self-similarity a gradual reweighted regularization is proposed for matrix completion and image restoration. Sparsity-promoting regularization produces much sparser representation of grouped nonlocal similar blocks of image by solving a nonconvex minimization problem. Moreover, an alternation direction method of multipliers algorithm is developed to speed up iterative solving of the above problem. Image block classification further enhances the adaptivity of the proposed method. Experiments on simulated matrix and natural image show that the proposed method obtains better image restoration results, where most lost information is reorcovered and few artifacts are produced.

Multi-script text versus non-text classification of regions in scene images

Text versus non-text region classification is an essential but difficult step in scene-image analysis due to the considerable shape complexity of text and background patterns. There exists a high probability of confusion between background elements and letter parts. This paper proposes a feature-based classification of image blocks using the color autocorrelation histogram (CAH) and the scale-invariant feature transform (SIFT) algorithm, yielding a combined scale and color-invariant feature suitable for scene-text classification. For the evaluation, features were extracted from different color spaces, applying color-histogram autocorrelation. The color features are adjoined with a SIFT descriptor. Parameter tuning is performed and evaluated. For the classification, a standard nearest-neighbor (1NN) and a support-vector machine (SVM) were compared. The proposed method appears to perform robustly and is especially suitable for Asian scripts such as Kannada and Thai, where urban scene-text fonts are characterized by a high curvature and salient color variations.

Application of Computer-Aided Diagnosis Technology in Brain Tumour Detection

Accurate segmentation of brain tumour means that surgeons accurately remove the tumour without damaging other healthy tissues. At present, due to the differences in human brains, the widely used manual brain tumour segmentation method cannot guarantee its accuracy and reliability. Therefore, it is of great social and practical significance to work out an automatic and accurate brain tumour segmentation method based on the computer-aided technology. This paper proposes a novel brain tumour segmentation method based on the deep learning model of stacked de-noising auto-coder. Firstly, by model training, it obtains the parameters of the deep learning network, and then it extracts high-level abstract features of the input image data through the network and uses these features to translate the segmentation of brain tumour to the classification of image blocks. Finally, this paper applies the proposed method for the MRI images of real brain tumour patients to carry out segmentation of brain tumours, and then compares it with the manual brain tumour segmentation method. The results show that the computer-aided brain tumour segmentation method is more effective and accurate and can provide reliable basis for the removal of brain tumours by surgeons without damaging normal tissues.

基于压缩感知思想的图像分块压缩与重构方法

Compressed sensing theory has great advantages on data acquisition, data storage and transmission, data analysis and processing. It has become a research hotspot in recent years. Considering the difference of information distribution in most image signals, at the encoding side, image entropy estimation and edge detection mechanism were used to calculate information content on the base of classification of image blocks. Then we sampled the image in two different angles: image blocks were classified according to the information content into smooth, transitional, and texture blocks, different measurement rates were used; According to the distribution characteristics of the information content, we use different sampling rate allocation strategies for sampling. At the decoding side, different linear operators were constructed to do linear reconstruction according to the different types of image blocks and further using improved iterative thresholding algorithm to remove block effect and noise. The experiments proved that the algorithm improved the quality of reconstructed image and reduced the reconstruction time at the same time and is also available for images with many textures and edges.

A fast classification based method for fractal image encoding

In the present paper a fast and efficient fractal image encoding method based on classification of image blocks is presented. Two parameters are used to sort image blocks into disjoint classes: the direction of the approximate first derivative and a normalized root mean square error of the fitting plane in the given block. With the help of these parameters the number of domain blocks examined for a range block is reduced dramatically, and thus, the classification results in a considerable acceleration of the encoding process, without loss of the reconstruction fidelity. The proposed method is compared to recently developed fast classification algorithms and a ‘No search algorithm’, and its rate-distortion performance under the same encoding time limit is proved to be better than that of the others.

Identification of Lesser Cornstalk Borer-attacked Maize Plants using Infrared Images

The lesser cornstalk borer (Elasmopalpus lignosellus) is a pest that damages the maize plants in the initial growing phase causing stand reduction that can result in yield decrease. The machine vision system could be an alternative for the development of site-specific management of this pest. The objective of this work was to develop and evaluate a machine vision algorithm for identifying maize plants attacked by lesser cornstalk borer based on colour infrared images. To develop the algorithm, images of 40 maize plants were taken on different days after emergency. The plants were grown in pots, and 25 of them were infested with lesser cornstalk borer larvae and 15 were left healthy. The algorithm had three stages: leaf identification, image block classification, and plant classification. In the leaf identification stage, the plant leaves were segmented by thresholding the normalised difference vegetation index image. For the block classification stage, different neural network architectures and block sizes were tested for identification of non-attacked and attacked plant image blocks. Then, in the plant classification stage, discriminating functions were used to classify the scene as either a healthy or an attacked plant. The algorithm performance was compared with the performance of four human experts by using the Kappa coefficient of agreement. The largest size of image block, 9 by 9 pixels, was chosen because of its less computational exigency and because its performance was not significantly different from the other tested block sizes. The algorithm performance was significantly better than just one human expert. The Kappa coefficients for the algorithm and the three best human experts were 63.0 and 49.7%, respectively. The overall accuracy of the algorithm and the best three human experts was 81.6 and 73.4%, respectively.

Storage- and entropy-constrained classified vector quantization

This paper presents a novel variable-rate classified vector quantizer (CVQ) design algorithm for image coding. The design algorithm, termed storage- and entropy-constrained classified vector quantization (SECCVQ) algorithm, is able to control the rate and the storage size of the CVQ. The algorithm allocates the rate and storage size available to each class of the VQ optimally so that the average distortion of the SECCVQ is minimized. Classification of image blocks is based on the edge orientation of each block in the wavelet domain. To reduce the arithmetic complexity of the CVQ, we employ a novel partial distance codeword search algorithm in the wavelet domain. Simulation results show that the SECCVQ enjoys a low average distortion, a low encoding complexity, a high visual perception quality, and is well-suited for very low bit rate image coding.

Image block classification using stochastic imagesegmentation

The authors propose a new image block classification method. The proposed algorithm incorporates image context into the classification via pixel-based segmentation. To obtain a segmented image they adopt the stochastic model-based unsupervised image segmentation algorithm. Since the block classifier considers the grey level distribution in the block, it can differentiate edges from textures. Also, since the segmentation is executed independently at each small block, a parallel processor can be applied to obtain a real-time block classification.