International Journal of Pattern Recognition and Artificial Intelligence

The authors hereby formally submit a request to retract the manuscript titled "Depth Edge and Structure Optimization-Based End-to-End Self-Supervised Stereo Matching", which was submitted to IJPRAI on May 19, 2023, by Wenbang Yang and co-authors. The reasons for this retraction are as follows:

1. Errors or deficiencies were discovered in the data during the research process, which affect the scientific rigor and accuracy of the paper.

2. Unresolved academic ethical issues have been identified and require further verification and revision.

3. Upon verification, certain parts of the manuscript were found to overlap with previously published works, necessitating a thorough review and revision.

Vehicle trajectories represent an essential information source in intelligent transportation systems. Prior trajectory completion models based on Automatic Number Plate Recognition (ANPR) data have typically depended on vehicle re-identification results or road network information or have employed static knowledge graphs to integrate the two information sources. However, these methods have not taken into account the implicit temporal characteristics of trajectories in ANPR data and have neglected individual vehicle preferences. To address this void, this study proposes a Temporal Knowledge Graph-based Vehicle Trajectory Completion Model (TKG-VTC). The model implementation comprises three stages: first, ANPR data are converted into a temporal trajectory knowledge graph; second, knowledge representation learning is conducted using nontemporal relations, a biased temporal regularizer and multivector embeddings to embed the knowledge on the graph; and finally, the embedded results are employed to perform link prediction for incomplete trajectories, thereby restoring vehicle trajectories in ANPR data. Through model evaluation metrics and dimensionality reduction experiments, TKG-VTC is observed to demonstrate the best performance in completing trajectories when compared to TComplEx, TNTComplEx, and TeLM. This research introduces an innovative application of employing temporal knowledge graphs for trajectory reconstruction, which eliminates dependence on vehicle re-identification and road network information in previous methodologies. This is advantageous for enhancing the performance and dependability of vehicle trajectory data in intelligent transportation systems, as well as facilitating the implementation of trajectory prediction, demand analysis, and accident warning applications.

Scene text style transfer without a language barrier is an open challenge for the video and scene text recognition community because this plays a vital role in poster, web design, augmenting character images, and editing characters to improve scene text recognition performance and usability. This work presents a new model, called Script Independent Scene Text Style Transfer Network (SISTSTNet), for extracting scene characters and transferring text style simultaneously. The SISTSTNet performs mapping in language-independent feature space for transferring style. It is designed based on a Style Parameter Network and Target Encoder Network through lightweight MobileNetv3 convolutional and residual blocks to capture the style and shape to generate target characters. Similarly, a generative model is explored through the Visual Geometry Group (VGG) network for character replacement. The SISTSTNet is flexible and works on different languages and arbitrary examples in a neat and unified fashion. The experimental results on images in various languages, namely, English, Chinese, Hindi, Russian, Japanese, Arabic, Greek, and Bengali and cross-language validation demonstrate the effectiveness of the proposed method. The performance of the method is superior compared to the state-of-the-art methods in terms of quality measures, language independence, shape-preserving, and efficiency. The code and dataset will be released to the public to support reproducibility.

In recent years, single image super-resolution (SISR) methods using convolutional neural networks (CNN) have achieved satisfactory performance. Nevertheless, the large model scale and the slow inference speed of these methods greatly limit the application scenarios. In this paper, we propose a two-stage three-dimensional attention network (ATTNet) for lightweight image super-resolution. First, we put forward the spatial feature encoder–decoder (SFE-D) with a spatial attention mechanism. Next, the channel transposed attention module (CTAM) with a channel self-attention mechanism is designed. Both the modules are used for fine feature extraction in the low-resolution stage. Finally, the content-based pixel recombination module (CPRM) is proposed to reconstruct the detailed content with a joint attention mechanism in the high-resolution stage. According to our experimental results, significant performance in terms of the quantitative metrics and the subjective visual quality can be achieved on average compared with the state-of-the-art lightweight SISR algorithms.

This paper presents a method for evaluating the quality of images altered by Gaussian blur. The method is based on the observation of bokeh mode images where the region of interest (foreground) is sharp, while the remaining parts (background) are intentionally blurred to enhance the perceptual quality of the image. The blurriness of the background increases attention towards the foreground part of the image. The proposed quality metric is obtained by combining the attention factor and the sharpness of the region of interest. The accuracy, in terms of Spearman’s-rank-order correlation-coefficient (SROCC), for popular and publicly available databases such as LIVE, VCL, TID2008, CSIQ, and TID2013, is 0.963, 0.925, 0.900, 0.930, and 0.930, respectively. The proposed method achieves high and consistent Spearman’s rank-order correlation coefficient (SROCC) values compared to the majority of state-of-the-art algorithms. Furthermore, in terms of speed, the proposed method surpasses other state-of-the-art methods. The MATLAB code of the proposed metric is publicly available at https://drive.google.com/drive/folders/1SRmUp0N157Ati9l3kV13uoCxw5PhMgQn?usp=sharing.

Watermarking has been developed as a beneficial approach to resolve security problems such as copyright defense, legitimate ownership, and authenticity of digital data. Grayscale or binary images and color images are the most important types of embedded watermarks in the existing video watermarking schemes, while dark frame watermarking is barely used. A new robust blind video watermarking system with an Advanced Squeezed Convoluted Ebola Neural Network (ASCENN) approach is suggested to secure copyright, reduce noise, and authenticate the authenticity of dark frames. The proposed method selects the dark frames from the extracted input video. Then the selected dark frames are embedded and extracted using Finite Ridgelet Transform (FRT), All Phase Discrete Cosine Biorthogonal Transform (APDCBT), and Singular Value Decomposition (SVD) techniques. The robustness of the introduced approach is established by evaluating it through various image watermarking techniques and evaluating its performance against geometric, nongeometric, and combinational attacks. While comparing with the other existing video watermarking approaches, the proposed method has obtained a maximum correlation coefficient of 1, and the signal-to-noise ratio of 75dB, at minimum error rate of 0.0215, respectively.

Aerial gaze target recognition is an important step in aerial eye control interaction. In order to achieve accurate aerial gaze target recognition, the VT4LM recognition algorithm was constructed in this paper. By this algorithm, the facial images containing the eyes of the flight operators were first inputted into Vision Transformer (ViT) to extract the local features, and the head posture of the flight operators was inputted into four LSTMs to extract the global features. Then, the local and global features of the flight operators were inputted into three fully connected layers, two dropout layers and one softmax classifier. Finally, the recognition result of aerial gaze target was obtained. In this paper the effectiveness of the VT4LM recognition algorithm was verified through the identification of four aerial line-of-sight gaze targets named Head-up display, Accelerator push rod, Control lever and Rudder by four flight operators during simulated flight. The experimental results showed that the accuracy of the VT4LM algorithm for aerial gaze target recognition reached 89.29% and the Cross Entropy Loss was 1.45. Compared to the other three recognition methods, the VT4LM algorithm had the highest recognition accuracy and minimum loss. When using the VT4LM algorithm to detect four simulated flight operators staring at four aerial gaze targets, the recognition accuracy was all higher than 85.00%. It could be seen that the VT4LM algorithm had a good performance in aerial gaze target recognition.

The key to improve the service effect of virtual digital human is the personality trait recognition technology of users. However, the complexity of Chinese and the sparsity of short texts limit the accuracy of personality trait recognition. Therefore, we applied emotional features to the research of personality trait recognition and constructed a deep learning model for Chinese corpus personality trait recognition to improve the accuracy of personality trait recognition. The model (${\text{R-B}}_e)$ extracted semantic features and embedded emotional features through RoBERTa-wwm-ext Chinese pre-training model, BiLSTM neural network, and attention mechanism. Some experiments were carried out using two Chinese datasets with different scenes. The comparative experiments and ablation experiments showed that the overall effect of the personality trait recognition model (${\text{R-B}}_e)$ is the best, and the Macro-F1 value reaches 75.3%, which is significantly better than the existing models. The results showed that emotional features can significantly improve the accuracy of multi-label personality trait classification in both Chinese daily life scenarios and online scenarios. The proposed model can integrate emotional features and Chinese semantic features to improve the accuracy of personality traits recognition. These findings support the development of personality trait recognition and the application of virtual digital human.

This paper proposes a few-shot pedestrian re-identification (Re-ID) model based on an improved ResNet50 with a compression and stimulation module, which is named CS-ResNet50. It combines the meta-learning framework with metric learning. This method first compresses residual network channels, then stimulates them to achieve the effect of feature weighting, ultimately making feature extraction more accurate. The research makes the model learn how to finish new tasks efficiently from its experience that it has obtained in the training process of former subtasks. In each subtask, the dataset is divided into a gallery set and a query set, where the model parameters are trained. In this way, the model can be trained efficiently and adopted to new tasks rapidly, which could solve few-shot Re-ID problems. Compared with the baseline, the proposed model improves two indicators efficiently on two Re-ID datasets and achieves better Re-ID effect in few-shot mode.

Detection transformers (DETR) have provided a novel solution to human–object interaction detection in a set-prediction manner, thanks to expressive learnable queries. However, few studies covered how queries implicitly affect model behaviors, which might contain clues for model improvement. Therefore, we propose two dataset-based analysis tools: score state space and query preference map. They provide data on the distribution of model predictions, which can reveal overall-level and query-level model properties. Starting with our baseline model, we find the model naturally regresses object boxes to overlap human boxes in the edge case of no-object verbs (stand, etc.), even without a related loss. We encourage this by patching the supervision with virtual objects, resulting in more stable query preference. Moreover, we show that two-stage decoders designed for cascade inference do not decouple tasks as intended. We infer this is caused by the empty instances used as negative samples, which suggests a redesign in the matching scheme. Further, we reveal how adding an oracle-query-based teacher model affects query roles with a tiny gain, indicating room for refinement. Our findings demonstrate how a simple focus on query behaviors can provide insights for improving models.

With the advent of deep neural networks, medical image analysis is able to predict results in advance in early detection and diagnosis of diseases found in the human body. Several deep neural network methodologies have been implemented for a quick and efficient analysis of medical images that detect and diagnose cancerous cell growth in any part of the human body. For improving the segmentation and classification accuracy, the paper has proposed a framework comprising modified DoubleU-Net for image segmentation and PolyNet architecture for image classification. The modified DoubleU-Net is composed of two U-Net architectures, in which U-Net1 makes use of ResNet-50 as an encoder in the place of VGG-16 (existing) and Atrous Spatial Pyramid Pooling (ASPP) is replaced by Waterfall Atrous Spatial Pooling (WASP) architecture in both U-Nets to improve the semantic image segmentation. For classifying the segmented medical images as benign or malignant, PolyNet architecture is implemented in the research. The research involves experiments on the brain tumor dataset and lung cancer dataset to analyze the performance of the proposed approach. The processing of the DoubleU-Net and modified DoubleU-Net is evaluated based on precision, Recall, Intersection over Union (IoU), and Dice Score as the performance metrics. Experimental findings indicate that the modified DoubleU-Net design outperformed the existing DoubleU-Net architecture in terms of performance parameters for segmentation. The efficiency of the PolyNet classifier has been evaluated against VGG-16 and Inception-V3 classifiers, in terms of accuracy, specificity, sensitivity, error rate, and computation time as the performance metrics. From the experimental results, it has been proved that the PolyNet classifier performs better than VGG-16 and Inception-V3 with improved accuracy, specificity, sensitivity, and computation time.