The radiological image investigation needs a lot of effort and expertise. This paper presents a semiautomatic tool designed for the detection of Pneumothorax and localization of Regions of Interest (ROIs) in X-Ray images through the utilization of UNet++ in an iterative manner. The evaluation of the Pneumothorax detection tool is conducted on a dataset introduced by Filice et al. The primary objective of this tool is to reduce the manual intervention required by radiologists, thereby enhancing overall throughput. In comparison to the state-of-the-art fastai model, our proposed model demonstrates a significant improvement, particularly in terms of recall. The study contributes to the field by introducing an efficient and effective tool for Pneumothorax detection, addressing the need for expeditious image analysis in radiology. The results suggest that the implemented UNet++ model outperforms existing methodologies, demonstrating its potential to revolutionize and optimize the radiological workflow. This research is pivotal for advancing the capabilities of automated tools in medical image analysis, ultimately benefiting healthcare professionals and patients alike.