Radiography remains the most performed exploration in radiodiagnosis services. Analyzing all radiographs performed in a health center is a huge workload for radiologist. As a consequence of this, a significant percentage of radiographs are not informed in many centers.
QUIBIM has developed a fully automated tool using Artificial Intelligence that analyzes Chest Radiographs and measures their probability of being abnormal. This tool helps radiologists to focus their efforts on potentially pathologic studies.
This Artificial Intelligence tool is able to estimate the probability of presence of fourteen different pathologies in Chest Radiographs. Once these probabilities are calculated the tool combines them to quantify the final abnormal probability of the images.
The tool provides visual information by means of heatmaps displayed over the analyzed radiographs. The heatmaps show the level of influence of each region of the images in the final abnormality score.
The probabilities and heatmaps provided by this tool can be used to optimize radiologist time by helping them to prioritize potentially pathologic radiographs.
- Pleural Thickening
Rajpurkar P, et al. CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning. CoRR. 2017;1711.05225.
Wang X, et al. ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases. IEEE CVPR. 2017;3462-3471
Simonyan K, et al. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv. 2015;1409.1556
Zhou B, et al. Learning Deep Features for Discriminative Localization. IEEE CVPR. 2016;2921–2929.