In this paper, we proposed a novel bonding technology to fabricate a microfluidic device based on Poly(methyl methacrylate) (PMMA). The method, which used chloroform and ethanol as miscible bonding solvent, can complete complex structures rapid assembly (10 min) at 40°C. A bonding strength of 267.5 N/cm² can be achieved, while the micro channel deformation was less than 7.26%. Then we utilized this method to produce a three layers micro mixer, which included a T-shaped inlet channel and six H-shaped mixing units. Numerical simulation indicated that, the well mixing length of the mixer was only about 6 mm when Re = 10. Finally, fluorescence microscopy was used to verify mixer performance. The method provided the potential for mass production of multilayer rigid polymer microfluidic devices.

A low noise interface ASIC for micro gyroscope with ball-disc rotor is realized in 0.5µm CMOS technology. The interface circuit utilizes a transimpedance pre-amplifier which reduces input noise. The proposed interface achieves 0.003°/s/Hz^1/2 noise density and 0.003°/s sensitivity with ±100°/s measure range. The functionality of the full circuit, including circuit analysis, noise analysis and measurement results, has been demonstrated.

With the increasing scale of e‐commerce, the complexity of image content makes commodity image classification face great challenges. Image feature extraction often determines the quality of the final classification results. At present, the image feature extraction part mainly includes the underlying visual feature and the intermediate semantic feature. The intermediate semantics of the image acts as a bridge between the underlying features and the advanced semantics of the image, which can make up for the semantic gap to a certain extent and has strong robustness. As a typical intermediate semantic representation method, the bag‐of‐visual‐words (BoVW) model has received extensive attention in image classification. However, the traditional BoVW model loses the location information of local features, and its local feature descriptors mainly focus on the texture shape information of local regions but lack the expression of color information. Therefore, in this paper, the improved bag‐of‐visual‐words model is presented, which contains three aspects of improvement: (1) multiscale local region extraction; (2) local feature description by speeded up robust features (SURF) and color vector angle histogram (CVAH); and (3) diagonal concentric rectangular pattern. Experimental results show that the three aspects of improvement to the BoVW model are complementary, while compared with the traditional BoVW and the BoVW adopting SURF + SPM, the classification accuracy of the improved BoVW is increased by 3.60% and 2.33%, respectively.

Polymer microchannels can be commonly processed using many non-lithographic methods for reducing the manufacturing cost and steps. In this research, an inexpensive and high-precision thermal engraving technology is developed and achieved to machine polymer microchannels ranging from tens to hundreds of micrometers. This paper presents the design of a thermal engraving device, the processing method and the experimental procedure. Thermal engraving microscribers can fabricate microchannels with a width less than 100 μm. Furthermore, the effects of velocity and temperature on the roughness of polymethyl methacrylate (PMMA) microchannels are also discussed. Finally, a smooth microchannel with these parameters optimally coordinated is achieved. Meanwhile, the contact angle (CA) and the electro-osmotic flow (EOF) of microchannels fabricated by this technology are also measured. The experimental results show that this method of fabrication has the advantages of low cost, high efficiency and small polymer microchannel size compared with several non-lithographic methods. This method of fabrication would be attractive for labs lacking extremely clean rooms and expensive photolithography apparatuses.

Based on the existing commodity image classification research, there is a lack of the use of the most recent color and texture features of the combination of commodity image classification. According to the local oppugnant color vector angle pattern (LOCVAP) feature extraction algorithm and the Deep Believe Network (DBN) classifier, the image classification is simulated and analyzed. The experimental results show that the results of LOCVAP descriptor have a greater improvement than those of other subcategories of feature descriptors.

Obtaining discriminative joint features is crucial for skeleton-based human action recognition. Current models mainly focus on the research of skeleton topology encoding. However, their predefined topology is the same and fixed for all action samples, making it challenging to obtain discriminative joint features. Although some studies have considered the complex non-natural connection relationships between joints, the existing methods cannot fully capture this complexity by using high-order adjacency matrices or adding trainable parameters and instead increase the computation parameters. Therefore, this study constructs a novel adaptive channel-enhanced graph convolution (ACE-GCN) model for human action recognition. The model generates similar and affinity attention maps by encoding channel attention in the input features. These maps are complementarily applied to the input feature map and graph topology, which can realize the refinement of joint features and construct an adaptive and non-shared channel-based adjacency matrix. This method of constructing the adjacency matrix improves the model’s capacity to capture intricate non-natural connections between joints, prevents the accumulation of unnecessary information, and minimizes the number of computational parameters. In addition, integrating the Edgeconv module into a multi-branch aggregation improves the model’s ability to aggregate different scale and temporal features. Ultimately, comprehensive experiments were carried out on NTU-RGB+D 60 and NTU-RGB+D 120, which are two substantial datasets. On the NTU RGB+D 60 dataset, the accuracy of human action recognition was 92% (X-Sub) and 96.3% (X-View). The model achieved an accuracy of 96.6% on the NW-UCLA dataset. The experimental results confirm that the ACE-GCN exhibits superior recognition accuracy and lower computing complexity compared to current methodologies.