Advanced Computer Vision Model for Lung Tumor Segmentation and Detection in Radiology and Histopathology Images

For effective treatment scheme and initial diagnosis, it is important to segue and detect the lung tumors.
Tumor variability, low contrast and overlapping tissue structure are some problems with traditional methods that make
it challenging to accurately and detect it. So, a newly advanced computer vision model that uses a Binary Gennet
Optimizer Tund Forward Neural Network (BGO-FFNN) method is used to sort tumors. The dataset contains anotate
CT scan and histopathology slides that were taken from publicly available repository. Gabber filters are used in
preprosauting to get rid of noise and strengthen the opposite, making it easier to see the edges of the tumor. Multi
scale edge-watthers' techniques work together for different tumor areas simultaneously. Local binary pattern (LBP)
is used to achieve texture features, which are important to explain the difference between a variety of tumors. BGO
is used to improve FFNN, and then classification performance is tested. The results showed that the suggested
model improves with recall (94.48%), F1-score (97.21%), accuracy (95.14%), and accurate (99.71%). Results
suggest that the proposed advanced computer vision model works better than standard people. This hybrid model
creates a major difference on how well the tumor is found and classified. It is very likely to use in the clinic and helps radiologists and pathologists assess and diagnose the exact tumor. These enrichment improves clinical accuracy
and streamlines the workflow in clinical settings. As technology is developed, further integration of artificial
intelligence in medical imaging can lead to more significant progress in patient results.