TY - GEN
T1 - Dual-Input Type Convolutional Neural Networks Employing Color Normalized and Nuclei Segmented Data for Histopathology Image Classification
AU - Demirel, Osman
AU - Akhtar, Muhammad Tahir
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Improvements in Convolutional Neural Network (CNN) have been widely successful for histopathology image classification. However, color normalization for data preprocessing and nuclei segmentation for feature extraction should also be considered for further performance boost, data redundancy elimination, and provision of distinguishing information. These techniques are known to improve generalizability. However, there is a need to find ways to use the data obtained from color normalized and segmented data for training. In this work, dual-input CNN (DiCNN), concatenated-input CNN (CiCNN), and ensemble CNN (ECNN) are trained and tested with color normalized and nuclei segmented data. The normalization technique is chosen based on correlation and structural similarity. The segmentation method is chosen based on the best-performing normalization technique for consistency and generalizability. The results show that normalized and segmented inputs results in better binary classification with CiCNN outperforming other methods. However, for multiclass classification raw data training is advantageous for all approaches.
AB - Improvements in Convolutional Neural Network (CNN) have been widely successful for histopathology image classification. However, color normalization for data preprocessing and nuclei segmentation for feature extraction should also be considered for further performance boost, data redundancy elimination, and provision of distinguishing information. These techniques are known to improve generalizability. However, there is a need to find ways to use the data obtained from color normalized and segmented data for training. In this work, dual-input CNN (DiCNN), concatenated-input CNN (CiCNN), and ensemble CNN (ECNN) are trained and tested with color normalized and nuclei segmented data. The normalization technique is chosen based on correlation and structural similarity. The segmentation method is chosen based on the best-performing normalization technique for consistency and generalizability. The results show that normalized and segmented inputs results in better binary classification with CiCNN outperforming other methods. However, for multiclass classification raw data training is advantageous for all approaches.
KW - Color Normalization
KW - Dual Input CNN
KW - Histopathology
KW - Image Classification
KW - Nuclei Segmentation
UR - http://www.scopus.com/inward/record.url?scp=85168854512&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85168854512&partnerID=8YFLogxK
U2 - 10.1109/SSP53291.2023.10208033
DO - 10.1109/SSP53291.2023.10208033
M3 - Conference contribution
AN - SCOPUS:85168854512
T3 - IEEE Workshop on Statistical Signal Processing Proceedings
SP - 378
EP - 382
BT - Proceedings of the 22nd IEEE Statistical Signal Processing Workshop, SSP 2023
PB - IEEE Computer Society
T2 - 22nd IEEE Statistical Signal Processing Workshop, SSP 2023
Y2 - 2 July 2023 through 5 July 2023
ER -