Computer Science and ICT

Permanent URI for this collectionhttps://repository.chuka.ac.ke/handle/chuka/430

Browse

Browsing Computer Science and ICT by Subject "CNN"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    The Effect of Adaptive Learning Rate on the Accuracy of Neural Networks
    ((IJACSA) International Journal of Advanced Computer Science and Applications, 2021) Jepkoech, Jennifer; Mugo, David Muchangi; Kenduiywo, Benson K.; Too, Edna Chebet
    Learning rates in gradient descent algorithms have significant effects especially on the accuracy of a Capsule Neural Network (CNN). Choosing an appropriate learning rate is still an issue to date. Many developers still have a problem in selecting a learning rate for CNN leading to low accuracies in classification. This gap motivated this study to assess the effect of learning rate on the accuracy of a developed (CNN). There are no predefined learning rates in CNN and therefore it is hard for researchers to know what learning rate will give good results. This work, therefore, focused on assessing the effect of learning rate on the accuracy of a CNN by using different learning rates and observing the best performance. The contribution of this work is to give an appropriate learning rate for CNNs to improve accuracy during classification. This work has assessed the effect of different learning rates and came up with the most appropriate learning rate for CNN plant leaf disease classification. Part of the images used in this work was from the PlantVillage dataset while others were from the Nepal database. The images were pre-processed then subjected to the original CNN model for classification. When the learning rate was 0.0001, the best performance was 99.4% on testing and 100% on training. When the learning rate was 0.00001, the highest performance was 97% on testing and 99.9% on training. The lowest performance observed was 81% accuracy on testing and 99% on training when the learning rate was 0.001. This work observed that CNN was able to achieve the highest accuracy with a learning rate of 0.0001. The best Convolutional Neural Network accuracy observed was 98% on testing and 100% on training when the learning rate was 0.0001.
  • Thumbnail Image
    Item
    Exploring Entropy Pruning Coupled with Capsule Neural Network (Caps net) For Leaf Disease Classification.
    (2022) Jepkoech, J.; Too, E. C.; Mugo, D.
    In an attempt to detect plant leaf diseases faster, accurately, and more efficiently, researchers have adopted deep learning methods using models like CNN and CapsNet with some success. However, there is a need for improvement as the current models are computationally expensive in terms of time and model complexity. Our work improves the Capsule Neural Network(CapsNet) model by adding convolutional layers to collect enough data and prune using the Entropy-based method. Entropy-based pruning reduces the number of characters in the model by getting rid of useless features and retaining only the useful features for detection. This considerably reduces the parameter size and reduces model complexities in terms of time and computation. We used F1, F5, and varying folds to assess accuracy in pruned mode against normal models. We tested our idea using 9080 images of tomatoes from PlantVillage and on three models, namely; ResNet-50, VGG-16, and CapsNet. CapsNet was the best among the pruned models with 98.9% followed by ResNet-50 with 93% and VGG-16 at 89.99%. We observed that pruning might be a Superior and less computationally expensive method than VGG-16 and ResNet-50. This implies that the accuracy of such models can be improved through the introduction of pruning.