---

ACCURATE DETECTION OF SARCOMA TISSUE FROM CHEST X-RAY IMAGES USING DEEP LEARNING FRAMEWORK

Tuhel Ahmed, Ali Hamza, Wahad Ur Rahman

---

Download PDF

---

Abstract

This study discusses the development and evaluation of advanced deep-learning applications aimed at detecting lung tumors in the lungs. Lung cancer is a leading cause of cancer-related deaths in the United Kingdom, accounting for approximately 20% of such fatalities and affecting about 35,000 people annually. Early detection is crucial for treating lung cancer. Research has shown that X-ray imaging is effective for screening, but interpreting the 2D medical images is challenging for humans and implementing them widely would put additional strain on already overburdened radiology departments. I have developed an innovative deep-learning method for automatically identifying lung nodules, which could indicate early-stage lung cancer. This approach shows promise in reducing the workload on human resources. The model was evaluated using a separate dataset and demonstrates performance comparable to the most advanced existing tools, with an average sensitivity of 82%. Additionally, I have devised a complementary innovation that leverages hierarchical connections to improve the efficiency of computer-aided detection tools for tasks such as nodule detection.

 

   Keywords— Deep learning, Lung cancer detection; Yolov5; X-rays