Human action recognition has attracted considerable research attention in the field of computer vision, especially for classroom environments. However, most relevant studies have focused on one specific behavior of students. Therefore, this paper proposes a student behavior recognition system based on skeleton pose estimation and person detection. First, consecutive frames captured with a classroom camera were used as the input images of the proposed system. Then, skeleton data were collected using the OpenPose framework. An error correction scheme was proposed based on the pose estimation and person detection techniques to decrease incorrect connections in the skeleton data. The preprocessed skeleton data were subsequently used to eliminate several joints that had a weak effect on behavior classification. Second, feature extraction was performed to generate feature vectors that represent human postures. The adopted features included normalized joint locations, joint distances, and bone angles. Finally, behavior classification was conducted to recognize student behaviors. A deep neural network was constructed to classify actions, and the proposed system was able to identify the number of students in a classroom. Moreover, a system prototype was implemented to verify the feasibility of the proposed system. The experimental results indicated that the proposed scheme outperformed the skeleton-based scheme in complex situations. The proposed system had a 15.15% higher average precision and 12.15% higher average recall than the skeleton-based scheme did.

When original aerial photographs are combined with deep learning to classify forest vegetation cover, these photographs are often hindered by the interlaced composition of complex backgrounds and vegetation types as well as the influence of different deep learning calculation processes, resulting in unpredictable training and test results. The purpose of this research is to evaluate (1) data preprocessing, (2) the number of classification targets, and (3) convolutional neural network (CNN) approaches combined with deep learning’s effects on high-resolution aerial photographs to identify forest and vegetation types. Data preprocessing is mainly composed of principal component analysis and content simplification (noise elimination). The number of classification targets is divided into 14 types of forest vegetation that are more complex and difficult to distinguish and seven types of forest vegetation that are simpler. We used CNN approaches to compare three CNN architectures: VGG19, ResNet50, and SegNet. This study found that the models had the best execution efficiency and classification accuracy after data preprocessing using principal component analysis. However, an increase in the number of classification targets significantly reduced the classification accuracy. The algorithm analysis showed that VGG19 achieved the best classification accuracy, but SegNet achieved the best performance and overall stability of relative convergence. This proves that data preprocessing helps identify forest and plant categories in aerial photographs with complex backgrounds. If combined with the appropriate CNN algorithm, these architectures will have great potential to replace high-cost on-site forestland surveys. At the end of this study, a user-friendly classification system for practical application is proposed, and its testing showed good results.

Studies on room monitoring have only focused on objects in a singular and uniform posture or low-density groups. Considering the wide use of convolutional neural networks for object detection, especially person detection, we use deep learning and perspective correction techniques to propose a room monitoring system that can detect persons with different motion states, high-density groups, and small-sized persons owing to the distance from the camera. This system uses consecutive frames from the monitoring camera as input images. Two approaches are used: perspective correction and person detection. First, perspective correction is used to transform an input image into a 2D top-view image. This allows users to observe the system more easily with different views (2D and 3D views). Second, the proposed person detection scheme combines the Mask region-based convolutional neural network (R-CNN) scheme and the tile technique for person detection, especially for detecting small-sized persons. All results are stored in a cloud database. Moreover, new person coordinates in 2D images are generated from the final bounding boxes and heat maps are created according to the 2D images; these enable users to examine the system quickly in different views. Additionally, a system prototype is developed to demonstrate the feasibility of the proposed system. Experimental results prove that our proposed system outperforms existing schemes in terms of accuracy, mean absolute error (MAE), and root mean squared error (RMSE).

In recent years, hardware advancements have enabled natural language processing tasks that were previously difficult to achieve due to their intense computing requirements. This study focuses on paraphrase generation, which entails rewriting a sentence using different words and sentence structures while preserving its original meaning. This increases sentence diversity, thereby improving the performance of downstream tasks, such as question–answering systems and machine translation. This study proposes a novel paraphrase generation model that combines the Transformer architecture with part-of-speech features, and this model is trained using a Chinese corpus. New features are incorporated to improve the performance of the Transformer architecture, and the pointer generation network is used when the training data contain low-frequency words. This allows the model to focus on input words with important information according to their attention distributions.