Search

The increasing interest in filter pruning of convolutional neural networks stems from its inherent ability to effectively compress and accelerate these networks. Currently, filter pruning is mainly divided into two schools: norm-based and relation-based. These methods aim to selectively remove the least important filters according to predefined rules. However, the limitations of these methods lie in the inadequate consideration of filter diversity and the impact of batch normalization (BN) layers on the input of the next layer, which may lead to performance degradation. To address the above limitations of norm-based and similarity-based methods, this study conducts empirical analyses to reveal their drawbacks and subsequently introduces a groundbreaking complex hybrid weighted pruning method. By evaluating the correlations and norms between individual filters, as well as the parameters of the BN layer, our method effectively identifies and prunes the most redundant filters in a robust manner, thereby avoiding significant decreases in network performance. We conducted comprehensive and direct pruning experiments on different depths of ResNet using publicly available image classification datasets, ImageNet and CIFAR-10. The results demonstrate the significant efficacy of our approach. In particular, when applied to the ResNet-50 on the ImageNet dataset, achieves a significant reduction of 53.5% in floating-point operations, with a performance loss of only 0.6%. [ABSTRACT FROM AUTHOR]

The evolution mechanism behind low cycle fatigue crack propagation behavior subject to overload was investigated by experimental and numerical methods. The retardation effect after the overload application was investigated by using compact tension (CT) specimens. An analysis model was established to consider accumulative plastic damage for CT specimen under single tensile overload. An elastic-plastic numerical model is established, and the stress-strain field near the crack tip is quantified by different numerical parameters. The crack opening displacement was got to reflect the influence of overload on plasticity-induced crack closure, and the evolution mechanism and accumulative plastic damage were investigated by exploring the residual compressive stress near the crack tip. The results show that the residual compressive stress produced by the tensile overload has great influence on the retardation effect, and the large plastic deformation near the crack front region after an overload witnesses a more significant crack closure phenomenon. The magnitude of residual compressive stress near crack tip increases with the increasing overload ratio, that sharpens the retardation effect after single tensile overload. [ABSTRACT FROM AUTHOR]

The recognition of fish species holds significant importance in aquaculture and marine biology. However, it is a challenging problem due to the high similarity among intra-genus species. Existing recognition methods primarily seek prominent features of the species. However, we believe that the diverse levels of similarity between a species and other species can also function as implicit characteristics for that specific species. Based on this perspective, we propose a dual-branch fusion network for fine-grained fish species recognition utilizing inter-species similarity. This approach consists of a backbone network and two branches for coarse- and fine-grained recognition. In the coarse-grained branch, we designed a guidance matrix and species similarity labels to facilitate the generation of species similarity information. In the fine-grained branch, features from the backbone network are fused with similarity information to achieve precise recognition. Finally, fine-tuning the neural network through loss functions. We conduct experimental validation on three publicly available fish datasets, yielding excellent accuracy outcomes. Code is available at https://github.com/xingxing317/fish_classification. [ABSTRACT FROM AUTHOR]

Purpose: To evaluate the changes in the liver volume and function after direct intrahepatic portocaval shunt (DIPS) in patients with Budd-Chiari syndrome (BCS) with diffuse hepatic vein (HV) occlusion. Methods: The clinical data of 29 patients with BCS who underwent DIPS for intractable ascites caused by diffuse hepatic vein occlusion in the Affiliated Hospital of Xuzhou Medical University were analysed retrospectively; the patients included 8 males and 21 females, with an average age of 33.3 ± 6.3 years. The patients underwent abdominal CT scanning and liver function examinations before DIPS, and 1 week, 3 months, 6 months and 12 months after DIPS. The changes of the liver volume and liver function before and after DIPS were compared. Results: All 29 patients underwent DIPS successfully. 28 patients survived during the follow-up of 12–33 months, with a median follow-up of 16 months. The patients' liver function were significantly improved at 3, 6 and 12 months after the operation compared to before the operation (P < 0.05). The liver volumes measured before the operation and 1 week, 3 months, 6 months and 12 months after the operation were 2124.586 ± 420.889 cm3, 1926.263 ± 372.268 cm3, 1480.592 ± 183.061 cm3, 1461.904 ± 153.027 cm3 and 1469.286 ± 148.549 cm3, respectively. Compared with the preoperative liver volume, the liver volume had decreased significantly at 1 week, 3 months, 6 months and 12 months after the operation (P < 0.05). However, there was no significant difference in the liver volumes at 6 and 12 months after the operation (P = 0.35). Conclusions: Direct intrahepatic portocaval shunt has achieved satisfactory clinical results in the treatment of BCS with diffuse hepatic vein occlusion. The congestive hepatomegaly was gradually reduced after the operation. The liver volume (which was defined as the clinical efficacy in this study) remained stable after 6 months. [ABSTRACT FROM AUTHOR]