Convolutional Neural Network (CNN) is popular deep learning framework with vast applications in image classification, segmentation, object detection etc., and has attracted attention of the machine learning community at large. In this publication, we aim to propose a model for classification of fruits. Our model is novel as it applies the concept of local connectivity of patterns in neural networks and learns low level features while preserving information about the geometry of objects and shapes. We demonstrated the effectiveness of our approach on a fruits dataset with 63 classes. The obtained results effectively demonstrate the local representation capacity of CNNs. We achieved test set accuracy of 96.63% and training set accuracy of 96.42%, which effectively exemplify the effectiveness of CNNs for this class of problems.
CITATION STYLE
Arshad, T., Jia, M., Guo, Q., Gu, X., & Liu, X. (2020). Fruit Classification Through Deep Learning: A Convolutional Neural Network Approach. In Lecture Notes in Electrical Engineering (Vol. 571 LNEE, pp. 2671–2677). Springer. https://doi.org/10.1007/978-981-13-9409-6_326
Mendeley helps you to discover research relevant for your work.