No abstract

Detection of the diseases on tomatoes in advance and making early intervention and treating increases the production amount, efficiency and quality which will satisfy the consumer with a more affordable shelf price. In this way, the efforts of the farmers who are waiting for the harvest throughout the season will not be wasted. In this paper a compact convolutional neural network (CNN) is proposed for diseases identification task where the network is comprised of only 6 layers that is why it is computationally cheap in terms of parameters employed in the network. This network is trained by using PlantVillage's tomato crops dataset which consists of 10 classes (9 diseases and 1 healthy). The proposed network is first compared with well-known pre-trained ImageNet deep networks using transfer learning approach. The results show that the proposed network performed better than pre-trained knowledge transferred deep network models and it is shown that there is no need to constitute very large, complicated network architectures to achieve a superior tomato diseases identification performance. Furthermore, to increase the performance of proposed network, data augmentation techniques are also employed during the network training. The proposed network achieves an accuracy, F 1 score, Matthews correlation coefficient, true positive rate and true negative rate of 99.70%, 98.49%, 98.31%, 98.49% and 99.81%, respectively using 9,077 unseen test images. Our results are better than or similar to the results of the state-of-the-art deep neural network approaches that used PlantVillage database and the proposed method employs the cheapest architecture.

Numerical optimization is one of the well-known problems in computer science. Day by day, new methods are developed by many researchers. Recently, optimization became an essential task for many disciplines, such as engineering, medicine, management and others. In many cases, optimization problems may require fast and efficient algorithms for real-time implementations. In this paper, a simple, fast and feasible algorithm is presented for the optimization of both uni-modal and multi-modal benchmark functions. A population based Bi-Attempted Based Optimization Algorithm (ABaOA) is a stochastic search method which searches a solution space with two fixed step-size displacement parameters and two mutation operators. The proposed algorithm is derived from Base Optimization Algorithm (BaOA) which uses basic arithmetic operations. The performance of ABaOA is tested on twenty well-known benchmark functions and the results are statistically compared with the seven well-known stochastic optimization algorithms. Three different statistical analyses were done on the results obtained from the ABaOA. Two non-parametric statistical comparisons with the mean values are performed by using Sign and Wilcoxon tests. The non-parametric statistical multiple comparisons of the proposed algorithm is performed by using the Friedman test. The non-parametric Friedman test of differences among repeated measures of these algorithms was conducted and referred a Chi-square value of 67.337, which was significant (p<0.05). Wilcoxon non-parametric pairwise comparison test was applied to analyze the difference of ABaOA statistically among the other algorithms. The test indicates that the introduced algorithm is statistically significant than other algorithms with a level of significance p < 0.05. The experimental results also show that the ABaOA is clearly superior to the compared stochastic optimization algorithms.

No abstract

No abstract