In our study, we propose a hybrid Convolutional Neural Network with Support Vector Machine (CNN-SVM) and Principal Component Analysis with support vector machine (PCA-SVM) methods for the classification of cocoa beans obtained by the fermentation of beans collected from cocoa pods after harvest. We also use a convolutional neural network (CNN) and support vector machine (SVM) for the classification operation. In the case of the hybrid model, we use a convolutional network as a feature extractor and the SVM is used to perform the classification operation. The use of PCA-SVM allowed for a reduction in image size while maintaining the main features still using the SVM classifier. Radial, linear and polynomial basis function kernels were used with various control parameters for the SVM, and optimizers such as the Stochastic Gradient Descent (SGD) algorithm, Adam, and RMSprop were used for the CNN softmax classifier. The results showed the robustness of the hybrid CNN-SVM model which obtained the best score with a value of 98.32% then the PCA-SVM based model had a score of 97.65% outperforming the standard CNN and SVM classification algorithms. Metrics such as accuracy, recall, F1 score, mean squared error (MSE), and MCC have allowed us to consolidate the results obtained from our different experiments.

— Plant identification is most often based on visual observations by botanists and systematists. Deep learning has become a tool that provides an alternative to automatic plant identification. Our study consists in implementing a method for plant recognition from herbarium specimens using deep learning classification methods. These methods were evaluated on the dataset of ten plant families from the national herbarium of Côte d'Ivoire. The proposed work uses CNN architectures such as DensNet-121, InceptionV3, VGG19, MobileNet, and ResNet101. The dataset contains 7543 images of herbarium specimens. The database is structured in three parts: training, testing, and validation. The accuracies obtained for the first scenario without preprocessing of herbarium specimen images are 76.94% for MobileNet, 77.77% for VGG19, and 77.96% for InceptionV3, 80.41% for ResNet101, and 83.47% for DensNet-121, respectively. The best performance was obtained with DensNet-121 with 83.47%. In the second scenario with preprocessing of herbarium specimens, the accuracies obtained were 82.80% for InceptionV3, 84.40% for VGG19, 85.53% for MobileNet, and 85.80% for ResNet101. The best accuracy was obtained with ResNet121 with 85.80%. From the analysis obtained, the results show that ResNet101 gives the best accuracy compared to the other architectures. In particular, the data preprocessing improves the prediction results, of the Convolutional Neural Network algorithms. Keywords— Deep learning, Herbarium specimens, image preprocessing, Convolutional Neural Network, Classification.

Biometric systems aim to reliably identify and authenticate an individual using physiological or behavioral characteristics. Traditional systems such as the use of access cards, passwords have shown limitations such as forgotten passwords, stolen cards, etc. As an alternative, biometric systems present themselves as efficient systems with a high reliability due to the physiological characteristics of each individual. This paper focuses on a deep learning method for fingerprint recognition. The described architecture uses a pre-processing phase in which grayscale images are represented on the RGB bands and then merged to obtain color images. On the obtained color images will be extracted the characteristics of the fingerprints textures.The fingerprint images after preprocessing are used in a deep convolution network system for decision making. The method is robust with an accuracy of over 99.43% and 99.53% with the respective variants densenet-201 and ResNet-50.

No abstract