Analyzing deep learning algorithms with statistical methods

Lucian Sergiu Trifan; Simona Moldovanu

doi:10.52846/stccj.2024.4.1.59

Authors

Lucian Sergiu Trifan Dunarea de Jos University of Galati
Simona Moldovanu Dunarea de Jos University

DOI:

https://doi.org/10.52846/stccj.2024.4.1.59

Keywords:

convolutional neural network, pretrained networks, skewness, kurtosis and standard deviation

Abstract

A number of farmers use human inspection to classify the quality of their fruits visually; this process could be done automatically with a Learning algorithm. This paper investigates the classification of apples, peaches, and oranges and verifies the quality of classification with the density of the edges detected in each image by the Sobel filter. The novelty of the study consists of comparing the classification accuracy of a convolutional neural network (CNN) and the statistical analysis of edge histograms. The edges are the main elements of what the CNN learns; this was the reason for verifying the deep learning algorithms with skewness, kurtosis, and standard deviation metrics extracted from the histogram of edge density detected by the Sobel filter. In the empirical process, four algorithms were proposed: a CNN with one convolutional layer, a DNN without convolutional layers, and two pre-trained networks, AlexNet and GoogLeNet. From clustering of the row data and analyzing the accuracy, the best classification is obtained by simple CNN for 20 epochs when the apples vs. oranges were classified.

References

S. Moldovanu and F. Batog, “The monitoring of burning buildings with convolutional neural network”, System Theory, Control and Computing Journal, vol. 3(2), pp. 1-8, (2023):

Y. LeCun, Y. Bengio, G. Hinton, Deep learning, Nature, vol. 521 pp. 436–444 (2015) 436.

C.T Cheng, T.Y. Ho, T.Y. Lee, C.C. Chang, C.C Chou, C.C. Chen, I. Chung and C.H. Liao, “Application of a deep learning algorithm for detection and visualization of hip fractures on plain pelvic radiographs”. Eur. Radiol. vol. 295, pp.469–5477, 2019.

R. Dandavate and V. Patodkar, “CNN and data augmentation-based fruit classification model”, 4th Int. Conf. on I-SMAC (IoT in Social, Mobile, Analytics and Cloud), pp. 784–787, 2020.

A. Kausar, M. Sharif, J. Park and D.R. Shin, “Pure-CNN: A framework for fruit images classification”, Int. Conf. on Computational Science and Computational Intelligence (CSCI) (2018), pp. 404-408

S.S.S. Palakodati, V.R Chirra, Y. Dasari and S. Bulla, “Fresh and Rotten Fruits Classification Using CNN and Transfer Learning”. Rev. d’Intelligence Artif. vol. 34, pp. 617–622, 2020.

M. Momeny, A. Jahanbakhshi, K. Jafarnezhad and Y.D. Zhang, “Accurate classification of cherry fruit using deep CNN based on hybrid pooling approach”. Postharvest Biol. Technol. vol. 166, pp. 111204, 2020.

U. Amin, M. I. Shahzad, A. Shahzad, M. Shahzad, U. Khan et al., “Automatic fruits freshness classification using CNN and transfer learning,” Appl. Sci., vol. 13(14), pp. 1–17, 2023.

H. Kang, and C. Chen, “Fruit detection, segmentation and 3D visualisation of environments in apple orchards”. Comput. Electron. Agric. vol. 171, pp. 105302, 2020,

M Miron, S Moldovanu, B.I. Ștefănescu, M. Culea, and S.M. Pavel, Culea-Florescu, A.L. “A new approach in detectability of microcalcifications in the placenta during pregnancy using textural features and k-nearest neighbors algorithm”. J. Imaging, vol. 8, pp. 81, 2022

T. Dewi, R. Rusdianasari, RD. Kusumanto, and S. Siproni, “Image Processing Application on Automatic Fruit Detection for Agriculture Industry,” Atlantis Highlights in Engineering, vol. 9, pp. 47-53, 2022.

A.R. Ali, J. Li and S.J. O’Shea, “Towards the automatic detection of skin lesion shape asymmetry, color variegation and diameter in dermoscopic images”. PLoS One, vol. 15, e0234352, 2020.

S. Moldovanu, L. Moraru and D. Bibicu “Characterization of myocardium muscle biostructure using first order features”, Dig. J. Nanomater Bios., vol. 6, pp. 1357-1365, 2011.

F.A. Damian, S. Moldovanu, N. Dey, A.S. Ashour and L. Moraru, “Feature Selection of Non-Dermoscopic Skin Lesion Images for Nevus and Melanoma Classification”. Computation.vol. 8, pp. 41, 2020.

T.Shanthi, and R.S. Sabeenian, “Modified Alexnet architecture for classification of diabetic retinopathy images”. Comput. Electr. Eng. vol. 76, pp. 56–64, 2019

Y. Cui, F. Zhou, J. Wang, X. Liu, Y. Lin and S. Belongie, “Kernel pooling for convolutional neural networks”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2921-2930, 2001.