Oracle bone inscriptions (OBIs) are a kind of hieroglyph, used about 3,600 years ago for divination and the recording of events. The characters on these OBIs are of great interest because they are precursors to the modern Chinese characters widely used across Asia today. However, as the OBIs were only discovered in 1899, there are currently few documents to describe them. Hence, recognizing and unlocking the meaning of OBIs helps to understand the ancient history of China, the evaluation of Chinese characters, and more. Currently, deep learning has made great progress and brought about a revolution in the research field of recognition, and shows good potential to meet the challenges of OBIs recognition. Due to the scarcity of resources, many OBIs contain only a small number of instances, which causes dataset imbalance and limits the accuracy of recognition. This paper attempts to provide a suite of OBIs recognition methods comprising an original OBIs dataset creation, dynamic dataset augmentation, and a novel deep learning-based recognition method. To this end, we create an original OBIs dataset and propose a modified Generative Adversarial Network for augmenting the original OBIs dataset. The augmented data is then dynamically selected for training the deep learning model considering the data imbalance problem. A novel model called C-A Net is proposed for OBIs recognition. The results of evaluation experiments show that the dynamical dataset augmentation can effectively locate a suitable training dataset for the deep learning model and solve the problem of imbalanced OBIs distribution. In addition, the recognition accuracy of C-A Net is 91.10%, which is higher than that of eight state-of-the-art models, and thus effectively suppresses the occurrence of overfitting. We also present an original OBIs dataset named OBI125, which is currently the only rubbing-type OBIs dataset that is open to the public. The code is available at http://www.ihpc.se.ritsumei.ac.jp/obidataset.html.

Deep neural networks (DNNs) have made significant achievements in a wide variety of domains. For the deep learning tasks, multiple excellent hardware platforms provide efficient solutions, including graphics processing units (GPUs), central processing units (CPUs), field programmable gate arrays (FPGAs), and application-specific integrated circuit (ASIC). Nonetheless, CPUs outperform other solutions including GPUs in many cases for the inference workload of DNNs with the support of various techniques, such as the high-performance libraries being the basic building blocks for DNNs. Thus, CPUs have been a preferred choice for DNN inference applications, particularly in the low-latency demand scenarios. However, the DNN inference efficiency remains a critical issue, especially when low latency is required under conditions with limited hardware resources, such as embedded systems. At the same time, the hardware features have not been fully exploited for DNNs and there is much room for improvement. To this end, this paper conducts a series of experiments to make a thorough study for the inference workload of prominent state-of-the-art DNN architectures on a single-instruction-multiple-data (SIMD) CPU platform, as well as with widely applicable scopes for multiple hardware platforms. The study goes into depth in DNNs: the CPU kernel-instruction level performance characteristics of DNNs including branches, branch prediction misses, cache misses, etc, and the underlying convolutional computing mechanism at the SIMD level; The thorough layer-wise time consumption details with potential time-cost bottlenecks; And the exhaustive dynamic activation sparsity with exact details on the redundancy of DNNs. The research provides researchers with comprehensive and insightful details, as well as crucial target areas for optimising and improving the efficiency of DNNs at both the hardware and software levels.

Due to the workforce shortage caused by the declining birth rate and aging population, robotics is one of the solutions to replace humans and overcome this urgent problem. This paper introduces a deep learning-based object detection algorithm for empty-dish recycling robots to automatically recycle dishes in restaurants and canteens, etc. In detail, a lightweight object detection model YOLO-GD (Ghost Net and Depthwise convolution) is proposed for detecting dishes in images such as cups, chopsticks, bowls, towels, etc., and an image processing-based catch point calculation is designed for extracting the catch point coordinates of the different-type dishes. The coordinates are used to recycle the target dishes by controlling the robot arm. Jetson Nano is equipped on the robot as a computer module, and the YOLO-GD model is also quantized by TensorRT for improving the performance. The experimental results demonstrate that the YOLO-GD model is only 1/5 size of the state-of-the-art model YOLOv4, and the $mAP$ of YOLO-GD achieves 97.38%, 3.41% higher than YOLOv4. After quantization, the YOLO-GD model decreases the inference time per image from 207.92 ms to 32.75 ms, and the $mAP$ is 97.42%, which is slightly higher than the model without quantization. Through the proposed image processing method, the catch points of various types of dishes are effectively extracted. The functions of empty-dish recycling are realized and will lead to further development toward practical use.

Recently, object detection witnessed vast progress with the rapid development of Convolutional Neural Networks (CNNs). However, object detection is mainly for multi-class tasks, and few networks are used to detect single-class multi-deformation objects. This paper aims to develop a lightweight object detection network for single-class multi-deformation objects to promote the practical application of object detection networks. First, we design a Densely Connected Multi-scale (DCM) module to augment the semantic information extraction of deformation objects. With the DCM module and other strategies incorporated, we design a lightweight backbone structure for object detection, namely, DCMNet. Then, we construct a lightweight Neck structure Ghost Multi-scale Feature (GMF) module for feature fusion using a feature linear generation strategy. Finally, with the DCMNet and GMF module, we propose the object detection network YOLO-SM for single-class multi-deformation objects. Extensive experiments demonstrate that our proposed backbone structure, DCMNet, significantly outperforms the state-of-the-art models. YOLO-SM achieves 97.66% mean Average Precision ( $mAP$ ) on the Barcode public dataset, which is higher than other state-of-the-art object detection models, and achieves an inference time of 55.45 frames per second (FPS), proving that the YOLO-SM has a good performance tradeoff between speed and accuracy in detecting single-class multi-deformation objects. Furthermore, in the single-class multi-deformation Crack public dataset, the $mAP$ of 86.11% is achieved, and an $mAP$ of 99.84% is obtained in the multi-class dataset Dish20, which is much higher than other state-of-the-art object detection models, proving that the YOLO-SM has good generalization ability.

The emergence of empty-dish recycling robots has alleviated problems such as labor shortages caused by an aging population. The detection and grasping of dishes play a crucial role in empty-dish recycling robots. However, due to the limited resources of edge devices, traditional object detection models require more space to store parameters and much computational overhead, limiting the development of empty-dish recycling robots. Therefore, this paper proposes an ultra-lightweight dish detection model YOLO-GS for an empty-dish recycling robot. We use the modified CSPDarknet as the backbone structure and design an ultra-lightweight Neck structure for efficient feature fusion. Meanwhile, we design a lightweight Head structure for object classification and bounding box coordinate regression by combining Ghost Shuffle Convolution (GSConv2D) and anchor-free method. For the empty-dish recycling robot to grasp the dishes, we design a dish grasp point extraction algorithm using image processing. Finally, TensorRT is used to optimize and accelerate the model for efficient and intelligent detection of dishes on the NVIDIA Jetson Xavier NX. The experimental results show that YOLO-GS achieves 99.380% mean Average Precision (mAP) with a parameter amount of 0.606 M . The inference speed of the TensorRT-optimized YOLO-GS algorithm reaches 31.371 frames per second, which meets the needs of real-time dish detection by the empty-dish recycling robot. The image of the empty-dish recycling robot demo is available at https://www.youtube.com/watch?v=pCBo1nzm3qU&t=22s.

Lower limb exoskeleton robots have shown significant research value due to their capabilities of providing assistance to wearers and improving physical motion functions. As a type of robotic technology, wearable robots are directly in contact with the wearer’s limbs during operation, necessitating a high level of human–robot collaboration to ensure safety and efficacy. Furthermore, gait prediction for the wearer, which helps to compensate for sensor delays and provide references for controller design, is crucial for improving the the human–robot collaboration capability. For gait prediction, the plantar force intrinsically reflects crucial gait patterns regardless of individual differences. To be exact, the plantar force encompasses a doubled three-axis force, which varies over time concerning the two feet, which also reflects the gait patterns indistinctly. In this paper, we developed a transformer-based neural network (TFSformer) comprising convolution and variational mode decomposition (VMD) to predict bilateral hip and knee joint angles utilizing the plantar pressure. Given the distinct information contained in the temporal and the force-space dimensions of plantar pressure, the encoder uses 1D convolution to obtain the integrated features in the two dimensions. As for the decoder, it utilizes a multi-channel attention mechanism to simultaneously focus on both dimensions and a deep multi-channel attention structure to reduce the computational and memory consumption. Furthermore, VMD is applied to networks to better distinguish the trends and changes in data. The model is trained and tested on a self-constructed dataset that consists of data from 35 volunteers. The experimental results show that FTSformer reduces the mean absolute error (MAE) up to $10.83\%$, $15.04\%$ and $8.05\%$ and the mean squared error (MSE) by $20.40\%$, $29.90\%$ and $12.60\%$ compared to the CNN model, the transformer model and the CNN transformer model, respectively.