Dermatoscopic image melanoma recognition based on CFLDnet fusion network

As a common skin disease, malignant melanoma has attracted great attention in dermatology with high morbidity and mortality. Traditional medical diagnoses rely on clinical experience and have a certain subjective bias. Aiming at the problem of similarity between classes of melanoma dermoscopic images and the imbalance of lesion data set, we propose a CFLDnet-based dermoscopic image lesion classification model, it mainly uses the improved convolutional block attention (CBA) DenseNet algorithm to enhance beneficial features. To better integrate the attention module into DenseNet without increasing too many parameters and wasting the computing resources, we discussed three types of variant networks derived from the CFLDnet. Moreover, to balance the different categories of the data set, we also use the Sample Focal Loss (SFL) to calculate the effective sample size of the data set and smooth the focal loss function. Large numbers of comparative experiments were done based on the ISIC2018 task3 dataset, the average recognition accuracy of the CFLDnet network proposed in this paper is 86.89%, which is much higher than other similar methods (VGG16, ResNet50, InceptionV3 and DenseNet121 with cross-entropy loss function).

[1]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[2]  Hitoshi Iyatomi,et al.  Skin lesion classification with ensemble of squeeze-and-excitation networks and semi-supervised learning , 2018, ArXiv.

[3]  Reda Kasmi,et al.  Classification of malignant melanoma and benign skin lesions: implementation of automatic ABCD rule , 2016, IET Image Process..

[4]  Muhammad Sharif,et al.  Construction of saliency map and hybrid set of features for efficient segmentation and classification of skin lesion , 2019, Microscopy research and technique.

[5]  Yang Song,et al.  Class-Balanced Loss Based on Effective Number of Samples , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Xin Zhang,et al.  Object class detection: A survey , 2013, CSUR.

[7]  Lars Schmidt-Thieme,et al.  Hyperparameter Optimization with Factorized Multilayer Perceptrons , 2015, ECML/PKDD.

[8]  Zellig S. Harris,et al.  Distributional Structure , 1954 .

[9]  Tanzila Saba,et al.  Skin lesion segmentation and classification: A unified framework of deep neural network features fusion and selection , 2019, Expert Syst. J. Knowl. Eng..

[10]  Waqar Mehmood,et al.  Breast Cancer Detection and Classification using Traditional Computer Vision Techniques: A Comprehensive Review. , 2020, Current medical imaging.

[11]  Russell C. Hardie,et al.  Skin Lesion Segmentation and Classification for ISIC 2018 by Combining Deep CNN and Handcrafted Features , 2019, ArXiv.

[12]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Muhammad Rashid,et al.  Microscopic melanoma detection and classification: A framework of pixel‐based fusion and multilevel features reduction , 2020, Microscopy research and technique.

[14]  Stephen W Dusza,et al.  The CASH (color, architecture, symmetry, and homogeneity) algorithm for dermoscopy. , 2007, Journal of the American Academy of Dermatology.

[15]  Sebastian Thrun,et al.  Dermatologist-level classification of skin cancer with deep neural networks , 2017, Nature.

[16]  Ross B. Girshick,et al.  Focal Loss for Dense Object Detection , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[17]  Alexander J. Smola,et al.  Stacked Attention Networks for Image Question Answering , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Rebecca L. Siegel Mph,et al.  Cancer statistics, 2018 , 2018 .

[19]  Yoshua Bengio,et al.  End-to-end attention-based large vocabulary speech recognition , 2015, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[20]  Yoshua Bengio,et al.  Neural Machine Translation by Jointly Learning to Align and Translate , 2014, ICLR.

[21]  R Marchesini,et al.  Optical imaging and automated melanoma detection: questions and answers , 2002, Melanoma research.

[22]  Muhammad Younus Javed,et al.  Melanoma Detection and Classification using Computerized Analysis of Dermoscopic Systems: A Review. , 2020, Current medical imaging.

[23]  Zhiguo Jiang,et al.  Application-Driven No-Reference Quality Assessment for Dermoscopy Images With Multiple Distortions , 2016, IEEE Transactions on Biomedical Engineering.

[24]  C. Boring,et al.  Cancer statistics, 1990 , 1990, CA: a cancer journal for clinicians.

[25]  J. Gershenwald,et al.  Staging and prognosis of cutaneous melanoma. , 2011, Surgical oncology clinics of North America.

[26]  Jie Chen,et al.  Classification and Research of Skin Lesions Based on Machine Learning , 2020, Computers, Materials & Continua.

[27]  Andrey Sorokin,et al.  Lesion Analysis and Diagnosis with Mask-RCNN. , 2018, 1807.05979.

[28]  Jagath C. Rajapakse,et al.  Deep neural network ensemble by data augmentation and bagging for skin lesion classification , 2018, ArXiv.

[29]  Tanzila Saba,et al.  Region Extraction and Classification of Skin Cancer: A Heterogeneous framework of Deep CNN Features Fusion and Reduction , 2019, Journal of Medical Systems.

[30]  Muhammad Sharif,et al.  Developed Newton-Raphson based deep features selection framework for skin lesion recognition , 2020, Pattern Recognit. Lett..

[31]  Muhammad Rashid,et al.  An integrated framework of skin lesion detection and recognition through saliency method and optimal deep neural network features selection , 2019, Neural Computing and Applications.

[32]  Koray Kavukcuoglu,et al.  Visual Attention , 2020, Computational Models for Cognitive Vision.

[33]  Yann Dauphin,et al.  Convolutional Sequence to Sequence Learning , 2017, ICML.

[34]  M. Emre Celebi,et al.  An Overview of Melanoma Detection in Dermoscopy Images Using Image Processing and Machine Learning , 2016, ArXiv.

[35]  Cordelia Schmid,et al.  Aggregating local descriptors into a compact image representation , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[36]  Jeffrey N. Rouder,et al.  Bayesian t tests for accepting and rejecting the null hypothesis , 2009, Psychonomic bulletin & review.

[37]  Hayit Greenspan,et al.  An Adversarial Learning Approach to Medical Image Synthesis for Lesion Detection , 2018, IEEE Journal of Biomedical and Health Informatics.

[38]  Thomas Mensink,et al.  Image Classification with the Fisher Vector: Theory and Practice , 2013, International Journal of Computer Vision.

[39]  Imran Ashraf,et al.  Review of Automated Computerized Methods for Brain Tumor Segmentation and Classification. , 2020, Current medical imaging.

[40]  In-So Kweon,et al.  CBAM: Convolutional Block Attention Module , 2018, ECCV.

[41]  S. Menzies,et al.  Frequency and morphologic characteristics of invasive melanomas lacking specific surface microscopic features. , 1996, Archives of dermatology.