Efficient Deep Spiking Multilayer Perceptrons With Multiplication-Free Inference.

Advancements in adapting deep convolution architectures for spiking neural networks (SNNs) have significantly enhanced image classification performance and reduced computational burdens. However, the inability of multiplication-free inference (MFI) to align with attention and transformer mechanisms, which are critical to superior performance on high-resolution vision tasks, imposes limitations on these gains. To address this, our research explores a new pathway, drawing inspiration from the progress made in multilayer perceptrons (MLPs). We propose an innovative spiking MLP architecture that uses batch normalization (BN) to retain MFI compatibility and introduce a spiking patch encoding (SPE) layer to enhance local feature extraction capabilities. As a result, we establish an efficient multistage spiking MLP network that blends effectively global receptive fields with local feature extraction for comprehensive spike-based computation. Without relying on pretraining or sophisticated SNN training techniques, our network secures a top-one accuracy of 66.39% on the ImageNet-1K dataset, surpassing the directly trained spiking ResNet-34 by 2.67%. Furthermore, we curtail computational costs, model parameters, and simulation steps. An expanded version of our network compares with the performance of the spiking VGG-16 network with a 71.64% top-one accuracy, all while operating with a model capacity 2.1 times smaller. Our findings highlight the potential of our deep SNN architecture in effectively integrating global and local learning abilities. Interestingly, the trained receptive field in our network mirrors the activity patterns of cortical cells.

[Effects of row spacing and sowing rate on vertical distribution of photosynthetically active radiation, biomass, and grain yield in winter wheat canopy.] / è¡Œè·å’Œæ’­ç§é‡å¯¹å†¬å°éº¦å† å±‚å ‰åˆæœ‰æ•ˆè¾å°„åž‚ç›´åˆ†å¸ƒã€ç”Ÿç‰©é‡å’Œç±½ç²’äº§é‡çš„å½±å“.

To clarify the effects of row spacing and sowing rate on the vertical distribution of canopy PAR, biomass, and grain yield in winter wheat, a field experiment was conducted without increa-sing water and fertilizer input. There were two row spacing modes, R1 (equal spacing, 20 cm+20 cm) and R2(wide and narrow row spacing, 12 cm+12 cm+12 cm+24 cm), and three sowing rates, D1 (low, 120 kg·hm-2), D2 (medium, 157.5 kg·hm-2), D3 (high, 195 kg·hm-2). The canopy photosynthetically active radiation (PAR) interception and utilization rate in different heights, population photosynthetic capacity, biomass, and grain yield were measured during the main growth stages of winter wheat. The results showed that both total PAR interception and upper layer PAR interception of winter wheat canopy under R1 treatment were significantly higher than those in R2 treatment, but those of the middle layer and lower layer were higher in R2 than in R1, and with significant difference in the middle layer. From flowering to maturity, the photosynthetic potential (LAD), population photosynthetic rate (CAP), PAR conversion rate, and utilization rate in R2 were all significantly higher than those in R1 under the same sowing rate, with the highest value under R2D2 treatment. With the increasing sowing rate, the population biomass (BA) and leaf biomass (BL) at different layers increased, but the individual biomass (BP) showed an opposite trend. Under the same sowing rate, BA, BL and BP in R2 were higher than that in R1 after the flowering stage. Among them, BA and BP had significant difference in row spacing treatments at the maturity stage, with significant difference between the two row spacing treatments being observed in BL of the middle and lower layers under D2 and D3 sowing rates. The spike number, grain number per spike, 1000-kernel weight, and grain yield of winter wheat among different treatments were the highest in R2D3, R2D1, R2D1, and R2D2, respectively. The 1000-kernel weight, grain number per spike and grain yield in R2 treatment were significantly higher than R1. In summary, the PAR interception in the middle and lower layers of winter wheat canopy was improved by changing row spacing, with positive consequence on the photosynthetic capacity of individual plant and population, PAR utilization and transformation efficiency, which finally increased biomass and grain yield. Therefore, optimizing the field structure and shaping the ideal population photosynthetic structure should pay more attention during the high-yield cultivation of winter wheat. Making full use of light resources per unit land area and excavating the photosynthetic production potential of crops were also critical to achieve high yield and efficiency. In this experiment, the population photosynthetic capacity, photosynthetic effective radiation utilization rate, and yield were the highest under the treatment of R2D2.

Wheat ear counting using K-means clustering segmentation and convolutional neural network.

BACKGROUND: Wheat yield is influenced by the number of ears per unit area, and manual counting has traditionally been used to estimate wheat yield. To realize rapid and accurate wheat ear counting, K-means clustering was used for the automatic segmentation of wheat ear images captured by hand-held devices. The segmented data set was constructed by creating four categories of image labels: non-wheat ear, one wheat ear, two wheat ears, and three wheat ears, which was then was sent into the convolution neural network (CNN) model for training and testing to reduce the complexity of the model. RESULTS: The recognition accuracy of non-wheat, one wheat, two wheat ears, and three wheat ears were 99.8, 97.5, 98.07, and 98.5%, respectively. The model R 2 reached 0.96, the root mean square error (RMSE) was 10.84 ears, the macro F1-score and micro F1-score both achieved 98.47%, and the best performance was observed during late grain-filling stage (R 2 = 0.99, RMSE = 3.24 ears). The model could also be applied to the UAV platform (R 2 = 0.97, RMSE = 9.47 ears). CONCLUSIONS: The classification of segmented images as opposed to target recognition not only reduces the workload of manual annotation but also improves significantly the efficiency and accuracy of wheat ear counting, thus meeting the requirements of wheat yield estimation in the field environment.