Multimode classification of caner using ultrasounds

Multimode classification of caner using ultrasounds

1. Introduction Existing Approaches/Related Works Problems in Existing Approaches Proposed Methodology Results and Discussion Conclusions and Future Work References Outline

2. ** Introduction to Cancer and proposed model **: Overview of Cancer, its factors and its treatment. Importance of Machine Learning in Cancer Detection using Ultrasounds . ** Challenges in Cancer Detection using Ultrasound images **: Lack of well- defined data. Difficult in handling large datasets. ** Proposed Methodology **: Utilization of Convolutional Neural Networks (CNNs) for Classification into different organs. Predicting the nature of cancer present in the organ (Gall Bladder, Breast). ** Results Compilation and Impact **: Diagnosing the nature of the cancer(malignant, benign and normal). Fastens the automates the process of diagnosing ultrasound. Introduction

3. 1. ** Multi-Organ Cancer Diagnosis via Deep Learning-Based Methods **: existing deep learning segmentation and classification methods for multi-organ cancer diagnosis using CNN based approach. 2. The existing approach for classifying breast cancers involves using ultrasound images, extracting features through statistical tests and shared data, transforming these features via a method called "watershed," and applying dimensionality reduction techniques like Principal Component Analysis (PCA) to create multidimensional sets of characteristics for classification. 3. ** Ultasound classification **: The framework for the algorithm applied image augmentation through flip, rotation, zoom, and contrast adjustment transformations at specified magnitudes. 4. ** Deep Learning-Based approach for abdominal ultrasound **: The existing approach for diagnosing gallbladder cancer (GBC) involves training, validating, and testing a deep learning model to automatically detect GBC using abdominal ultrasound data, showing comparable diagnostic performance to radiologists across various patient scenarios. Existing Approaches/Related Works

4. 1. ** Compelxity of Data **: - Difficulty in accurately distinguishing between bird calls and other environmental sounds, which could lead to false positives or negatives. 2. ** Complexity of image and automation **: - The complexity arises from the intricate nature of breast tissues and the variations in how they appear under ultrasound, which can make it hard to differentiate between benign and malignant lesions accurately . 3. ****: - The method relies on unsupervised learning, which can be less effective than supervised learning in learning to separate specific sounds from mixtures. 4. ** Limited Applicability **: - The primary issue with current deep learning models for diagnosing gallbladder cancer is their limited generalizability across different clinical settings and the trade-off between high sensitivity and specificity, impacting reliable clinical decision-making Problems in Existing Approaches

5. ** Ultrasound Detection **: Description: Detect whether the image provided is ultrasound or not Method: Classification of images into ultrasound or not using ResNet-50 **Organ Classifica tion **: Description: Classifying ultrasound image into the type of organ. Method: Non-overlapping Batch Partitioning and Translation to Spectrograms for Each Batch **Gall/Breast Cancer detection **: Description: Detecting Cancer in Breast and Gall Bladder and the nature of Cancer( Benign, Malignant,Normal ). Method: Predicting Cancer in the respective organ using CNN methods.VGG16 for Breast and ResNet50 for Gall Blader. ** Predictor Model **: Description: Compiling all the models together to work efficiently. Method: Constructing a pipeline for the models to work together in a systematic order. Proposed Methodology

6. Results and Discussion 1. **Ultrasound or not Model**: - Training Accuracy – 98.92% - Testing Accuracy- 98.86% - F1 Score- 0.98 2. **Organ Classifier Model**: - Training Accuracy – 99.50% - Testing Accuracy- 98.15% - F1 Score- 0.981 3. **Breast Cancer Detection Model**: - Training Accuracy – 90±3% % - Testing Accuracy- 96.86% - F1 Score- 0.95 4. **Gall Bladder Detection Model**: - Training Accuracy – 97.92% - Testing Accuracy- 96.86% - F1 Score- 0.93

7. Conclusions and Future Work **Utilization of Convolutional Neural Networks (CNNs)**: Pioneering strategy employing CNNs for diagnosis ultrasound for cancer detection. Focus on detecting cancer and classifying its nature(Benign, Malignant,Normal ), offering a robust identification method. **Achieving High Accuracy**: Attainment of exceptional 95% accuracy through supervised learning and ultrasonography-based feature extraction. Simplifies data collection and enhances understanding of cancer and organ classification. **Versatility in Sonography Diagnosis**: Demonstrated adaptability and versatility of the model hint at broader applications in ultrasound classification. Holds promise for automating ultrasound diagnosis. **Future Directions**: Expanding Model Capabilities: Future endeavors may involve extending the model to recognize and classify ailments with X-rays or different graphs. Development of a multi-label classification approach for comprehensive assessments of different organs.

8. References 1. Bailar, J. C., & Gornik , H. L. (1997). Cancer undefeated. The New England Journal of Medicine, 336(22), 1569–1574. https://doi.org/10.1056/nejm199705293362206 2. Zugazagoitia , J., Guedes, C., Ponce, S., Ferrer, I., Molina-Pinelo, S., & Paz-Ares, L. (2016). Current challenges in cancer treatment. Clinical Therapeutics, 38(7), https://doi.org/10.1016/j.clinthera.2016.03.026 1551–1566 . 3. Ali, S.; Li, J.; Pei, Y.; Khurram, R.; Rehman, K.u .; Rasool, A.B. State-of-the-Art Challenges and Perspectives in Multi-Organ Cancer Diagnosis via Deep Learning-Based Methods. Cancers 2021, 13, 5546. 4. C. Muñoz-Meza and Wilfrido Gómez. 2013. A feature selection methodology for breast ultrasound classification. In Proceedings of the 10th International Conference on Electrical Engineering, Computing Science, and Automatic Control (CCE’13). IEEE, Los Alamitos, CA. https://doi.org/10.1109/ICEEE.2013.6676056 [5] Xiao Zheng, Xie Faqiang , Shi Jianwei , and Niu Xiaolan Tang. 2020. Breast ultrasound image classification and seg mentation using convolutional neural networks. Advances in Multimedia Information Processing 2018 (2020), https://doi.org/10.1007/978-3-030-00764-5_19 200–211. [6] ALEXANDER Rakhlin , Alexey Shvets, Vladimir Iglovikov , and Alexandr A. Kalinin. 2018. Deep convolutional neural networks for breast cancer histology image analysis. arXiv:1802.00752. [7] Scotty Kwok. 2018. Multiclassaclassification ofabreast cancer in whole-slide images. In Proceedings of the 15th International Conference on Image Analysis and Recognition (ICIAR’18). IEEE, Los Alamitos, CA, 931–940. https://doi.org/10.1007/978-3-319 93000-8_106 [8] Yeeleng S. Vang, Zhen Chen, and Xiaohui Xie. 2018. Deep learning framework for multi-class breast cancer histology image classification. arXiv:1802.00931 [9] Yi Wang, Eun Jung Choi, Younhee Choi, Hao Zhang, Gong Yong Jin, and Seok-Bum Ko. 2020. Breast cancer classification in automated breast ultrasound using multiview convolutional neural network with transfer learning. Ultrasound in Medicine and Biology 2020 (May 2020), https://doi.org/10.1016/j.ultrasmedbio.2020.01.001 1

9. References [10] S. Basu, M. Gupta, P. Rana, P. Gupta and C. Arora, "Surpassing the Human Accuracy: Detecting Gallbladder Cancer from USG Images with Curriculum Learning," in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 2022 pp. 20854-20864. doi : 10.1109/CVPR52688.2022.02022 [11] Deep- learningaenabled ultrasoundabased detection ofagallbladder cancer in northernaIndia : a prospectiveadiagnostic study,2023 / The Lancet RegionalaHealth - Southeast Asia, ISSN: 2772-3682, Page: 100279 [12] ActivationaFunctions in NeuralaNetworks [12 Types & Use Cases]. (n.d.). ActivationaFunctions in NeuralaNetworks [12 Types & Use Cases]. https://www.v7labs.com/blog/neural-networks activation-functions. [13] Introduction to Convolution Neural Network - GeeksforGeeks . (2017, August 21). GeeksforGeeks . https://www.geeksforgeeks.org/introduction-convolution-neural network/ [14] D. (2022, June 1). Convolutional Neural Networks (CNN) —  Architectures Explained. Medium. https://medium.com/@draj0718/convolutional-neural-networks cnn-architectures-explained-716fb197b243 [15] Vimal, Vrince , et al. "Artificial intelligence‐based novel scheme for location area planning in cellular networks." Computational Intelligence 37.3 (2021): 1338-1354.