Ekstraksi Fitur Citra Grayscale dengan Convolutional Neural Networks

This study aims to explore the use of Convolutional Neural Networks (CNN) in feature extraction from grayscale images for avocado object identification. The process begins with taking a grayscale image of the avocado object to be recognized. Convolution is applied using a 3x3 horizontal Sobel kernel filter with a stride of 1 to the right, and a ReLU (Rectified Linear Unit) activation function to improve the network's ability to extract relevant features. After the convolution stage, pooling is carried out using the max pooling method to reduce the image dimension while retaining important information, thereby speeding up the training process and reducing the risk of overfitting. The processed image is then flattened to produce a feature vector that is ready to be used in classification. The results of the study indicate that the CNN approach can be used as an effective method for feature extraction and edge detection on avocado objects from grayscale images.