International Journal of Pattern Recognition and Artificial Intelligence

Cursive handwriting recognition (CHWR) is an interesting area of research as it has a wide range of applications but lacks an accurate approach to provide better results due to its character shapes, the non-uniform spacing between words and within a word, diverse placements of dots, and diacritics, and very low inter-class variation among individual classes. A novel CHWR model is proposed to enhance the recognition accuracy with high global stability. The proposed model introduces three major phases: pre-processing, feature extraction and classification. In the pre-processing stage, the noise removal and binarization are adapted with the intrusion of improved adaptive wiener filtering (IAWF) and structural symmetric pixels. A hybrid deep direction distribution feature extraction (HDDDFE) approach is proposed to extract directional Local gradient histogram (LGH), column gradient histogram (CGH) features and a wavelet convolutional neural network with Block Attention Module (WCNN-BAM) is proposed to extract deep global features (GF), profile features (PF) and dynamic features (DF). A novel double hidden layer gated recurrent neural network with a feature attention mechanism (ODHL-GRNN-FAM) is proposed to offer handwritten classification results. The developed model is evaluated with the IAM database and attains an overall recognition accuracy of 98%, precision of 97%, f-measure of 97.99%, character error rate (CER) of 1.23%, word error rate (WER) of 4.8%, respectively.

Shearer drum automatic height adjustment strategy under mining environment is based on the recognition of coal–rock interface and the ground penetrating radar (GPR) was used for coal–rock interface recognition in the study. First, a model was built to study the radar echo in complex coal seam and some simulations were made to study the influence of radar parameters. Second, the experiment study was implemented in the coal mine working face in Tengzhou city, Shandong province, China. In this study, it was applied for radar image creation, including the start time correction, filtering technique, Hilbert transform, A-scan, and B-scan. The support vector machine (SVM) method was used for searching the coal–rock interface echo in lots of waveforms. The coal–rock interface could be found clearly and intuitively in the radar images by the above method in unknown complex coal seam structure and the error is less than 2% in A-scan mode. The results show that the method can stably and reliably find the coal–rock interface even in dynamic scenarios with the accuracy of 95%, where the root mean square error (RMSE) is and the 0.1. The radar antenna can be fixed to the shearer rocker arm in real time during mining to detect the thickness of coal seam in looking-ahead, top/bottom and shear moving direction.

Wireless image sensor networks (WISNs) collect surveillance images, resulting in copious quantities of data requiring processing and transmission within the network. To reduce and balance energy expenditure during in-network image data processing and transmission, this study introduces an energy-saving strategy based on image super-resolution for WISNs assisted by cloud. The strategy constructs an image data processing and transmission system for WISNs, leveraging cloud infrastructure. By utilizing image quality feedback, the adaptive adjustment of image node sensing resolution is facilitated, minimizing image data transmission volume while ensuring high-quality super-resolution reconstructed image. The establishment of low-overhead, multipath transmission routes enables image nodes to transmit data in blocks, contingent upon neighboring transmission node statuses, thereby equalizing energy consumption for image data transmission. Cloud servers are employed to reconstruct high-resolution images and provide users with surveillance images that satisfy quality requisites. Empirical findings demonstrate that the proposed approach extends the lifetime of WISNs, furnishes users with superior-quality surveillance images, and markedly enhances network monitoring efficacy.

In recent years, super-resolution neural networks have achieved good results in restoring super-resolution images from low-resolution ones. However, most subsequent tasks based on super-resolution images such as object detection are done by the computer. Considering this situation, we propose a learnable downsampler–upsampler pair, which can realize both the downscaling process and the upscaling process by neural networks, and is jointly trained with the YOLOV5 network to optimize the object detection task. Thus, different from existing super-resolution networks, the entire downsampler–upsampler pair is optimized for machine perception. In addition, to further reduce the size of the downsampled images, we also propose a differentiable method for estimating image entropy and add it to the loss function. We verify the effectiveness of our method on the pothole dataset and use scale factors 2× and 4× to prove that our method is capable of diverse resizing levels. The experimental results show that using our learnable downsampler–upsampler pair as a resizing method can highly improve the detection performance compared with other resizing techniques.

Many different histogram equalization (HE)-based image enhancement methods have been developed to overcome the problems of low or high image brightness, contrast sensitivity, and difficulty in revealing details of dark areas under low-light environments. In this paper, a novel image enhancement method based on HE and adaptive gamma correction with weight distribution (AGCWD) is proposed for natural and effective image enhancement. In this method, histogram stretching is performed on Red–Green–Blue (RGB) color components of image, and then the color space of RGB image is converted to Hue–Saturation–Intensity (HSI) color space. The histograms of S and I components are divided into sub-histograms according to the exposure threshold. The underexposure regions are enhanced with a new AGCWD. Then, the color space of HSI image is converted to RGB color space. Finally, the HE is applied to the input image with the obtained image histogram map. Thus, the method has not only effectively prevented the over-enhancement of the contrast but also obtained the quality and natural enhanced image. The proposed method is compared with the most known contrast enhancement methods and low-light enhancement methods. Experimental results have supported that the proposed method outperforms other methods in terms of both visual perception and objective evaluation.

During the production, transportation and storage of nonwoven fabric mask, there are many damages caused by human or nonhuman factors. Therefore, checking the defects of nonwoven fabric mask in a timely manner to ensure the reliability and integrity, which plays a positive role in the safe use of nonwoven fabric mask. At present, the wide application of machine vision technology provides a technical mean for the defect detection of nonwoven fabric mask. On the basis of the pre-treatment of the defect images, it can effectively simulate the contour fluctuation grading and gray value change of the defect images, which is helpful to realize the segmentation, classification and recognition of nonwoven fabric mask defect features. First, in order to accurately obtain the image information of the nonwoven fabric mask, the binocular vision calibration method of the defect detection system is discussed. On this basis, the defect detection mechanism of the nonwoven fabric mask is analyzed, and the model of image processing based on spatial domain and Hough transform is established, respectively. The original image of the nonwoven fabric mask is processed by region processing and edge extraction. Second, the defect detection algorithm of nonwoven fabric mask is established and the detection process is designed. Finally, a fast defect detection system for nonwoven fabric mask is designed, and the effectiveness of the detection method for nonwoven fabric mask is analyzed with an example. The results show that this detection method has positive engineering significance for improving the detection efficiency of defects in nonwoven fabric mask.

The video moment retrieval task aims to fetch a target moment in an untrimmed video, which best matches the semantics of a sentence query. Existing methods mainly focus on utilizing two separate modules: one learns intra-modal relations to understand video and query contents, and the other explores inter-modal interactions to build a semantic bridge between video and language. However, intra-modal relations information can be easily overlooked when capturing inter-modal interactions. In fact, intra-modal relations and inter-modal interactions can be learned simultaneously within a unified module to make video and sentence guide each other. Towards this end, we propose a Cross-Modal Interaction Network (CMIN) for video moment retrieval by jointly exploring the intra-modal relations and inter-modal interactions between video frames and query words. In CMIN, a query-guided channel attention module is designed to suppress query-irrelevant visual features and enhance crucial contents; then a cross-attention module simultaneously considers intra-modal relations within each modality and fine-grained inter-modal interactions between frames and words, to enhance the semantic relevance between video and sentence query. Compared to the state-of-the-art methods, the experiments on two public datasets (Charades-STA and TACoS) demonstrate the superiority of our method.

Facial Expression Recognition (FER) is a prominent research area in Computer Vision and Artificial Intelligence that has been playing a crucial role in human–computer interaction. The existing FER system focuses on spatial features for identifying the emotion, which suffers when recognizing emotions from a dynamic sequence of facial expressions in real time. Deep learning techniques based on the fusion of convolutional neural networks (CNN) and long short-term memory (LSTM) are presented in this paper for recognizing emotion and identifying the relationship between the sequence of facial expressions. In this approach, a hyperparameter tweaked VGG-19 skeleton is employed to extract the spatial features automatically from a sequence of images, which avoids the shortcoming of the conventional feature extraction methods. Second, these features are given into bidirectional LSTM (Bi-LSTM) for extracting spatiotemporal features of time series in two directions, which recognize emotion from a sequence of expressions. The proposed method’s performance is evaluated using the CK+ benchmark as well as an in-house dataset captured from the designed IoT kit. Finally, this approach has been verified through hold-out cross-validation techniques. The proposed techniques show an accuracy of 0.92% on CK+, and 0.84% on the in-house dataset. The experimental results reveal that the proposed method outperforms compared to baseline methods and state-of-the-art approaches. Furthermore, precision, recall, F1-score, and ROC curve metrics have been used to evaluate the performance of the proposed system.

The modeling of channel and temporal information is of crucial importance for action recognition tasks. To build a high-performance action recognition network by effectively capturing channel and temporal information, we propose CLS-Net: an action recognition algorithm based on channel-temporal information modeling. The proposed CLS-Net characterizes channel and temporal information by inserting multiple modules to an end-to-end backbone network, including a channel attention module (CA module) for modeling channel information, a long-term temporal module (LT module) and a short-term temporal module (ST module) for modeling temporal information. Specifically, the CA module extracts the correlation between feature channels so the network can learn to selectively strengthen the features containing useful information and suppress the useless features through global information. The LT module moves some channels in the temporal dimension to realize information interaction across time domains and model global temporal information. The ST module enhances the motion-sensitive features by calculating the feature-level frame difference information and realizes the representation of local motion information. Since the multi-module insertion mode directly affects the whole model’s final performance, we propose a novel multi-module insertion mode instead of a simple series or parallel connection to ensure that the multiple modules can complement one another and cooperate with each other more efficiently. CLS-Net achieves SOTA performance on the EgoGesture and Jester dataset in the same type of network and achieves competitive results on the Something-Something V2 dataset.

When a top-down method is taken to the task of human pose estimation, the accuracy of joint point localization is often limited by the accuracy of human detection. In addition, conventional algorithms commonly encode the image to generate a heat map before processing, but the systematic error in decoding the heat map back to the original image has an impact on the positioning. Therefore, to address the two problems, we propose an algorithm that uses multiple angle models to generate the human boxes and then performs lightweight decoding to recover the image. The new boxes can better fit humans and the recovery error can be reduced. First, we split the backbone network into three sub-networks, the first sub-network is responsible for generating the original human box, the second sub-network is responsible for generating a coarse pose estimation in the boxes, and the third sub-network is responsible for a high-precision pose estimation. In order to make the human box fit the human body better, with only a small number of interfering pixels inside the box, models of the human boxes with multiple rotation angles are generated. The results from the second sub-network are used to select the best human box. Using this human box as input to the third sub-network can significantly improve the accuracy of the pose estimation. Then to reduce the errors arising from image decoding, we propose a lightweight unbiased decoding strategy that differs from traditional methods by combining multiple possible offsets to select the direction and size of the final offset. On the MPII dataset and the COCO dataset, we compare the proposed algorithm with 11 state-of-the-art algorithms. The experimental results show that the algorithm achieves a large improvement in accuracy for a wide range of image sizes and different metrics.

Accurate tumor segmentation in medical images plays an important role in clinical diagnosis and disease analysis. However, medical images usually have great complexity, such as low contrast of computed tomography (CT) or low spatial resolution of positron emission tomography (PET). In the actual radiotherapy plan, multimodal imaging technology, such as PET/CT, is often used. PET images provide basic metabolic information and CT images provide anatomical details. In this paper, we propose a 3D PET/CT tumor co-segmentation framework based on active contour model. First, a new edge stop function (ESF) based on PET image and CT image is defined, which combines the grayscale standard deviation information of the image and is more effective for blurry medical image edges. Second, we propose a background subtraction model to solve the problem of uneven grayscale level in medical images. Apart from that, the calculation format adopts the level set algorithm based on the additive operator splitting (AOS) format. The solution is unconditionally stable and eliminates the dependence on time step size. Experimental results on a dataset of 50 pairs of PET/CT images of non-small cell lung cancer patients show that the proposed method has a good performance for tumor segmentation.

Currently, evaluating the quality of strong-flavor Baijiu (SFB) heavily relies on subjective sensory analysis, resulting in large deviations in evaluation. However, as there are no existing evaluation criteria for SFB quality, this study aimed to extract trace components and design an evaluation model using gas chromatography–mass spectrometry (GC–MS). First, the key component data was analyzed using principal component analysis (PCA) and sparse principal component analysis (SPCA) to identify the most important principal components that represent the SFB samples. Second, KNN, DT, SVM, and BP analyses were then employed on the principal component data to determine the grade of the SFB samples. Finally, a price prediction model based on SPCA+BP was established to objectively evaluate the quality and price of SFB. The experimental results show that the proposed method can effectively realize the distinction and price prediction of SFB.

The accuracy and reliability of continuous space curve estimation is the key to global exploration. An improved artificial intelligence algorithm is proposed for continuous space analysis. First, a wavelet basis ANN algorithm is proposed to determine the discretization strategy in continuous space. The hidden layer node transfer function in a BP neural network is replaced by a wavelet basis function, and the modified BP neural network is composed of a wavelet neural network. Second, an improved wolf algorithm is established. The core wolf system ensures the precision of the whole exploration task. Finally, main and auxiliary double cores and a five-class decision factor are used to establish a population classification model to solve the convergence of the algorithm.