User login
Researchers around the world may be able to teach computers how to better detect and diagnose disease, thanks to > 100,000 chest x-ray images and corresponding data recently released by the NIH Clinical Center.
Reading and diagnosing chest x-rays requires careful observation, as well as knowledge of anatomy, physiology, and pathology. When that is combined with the need to consider all common thoracic diseases, it becomes hard to automate a consistent technique for reading images, the NIH says. With the free dataset, the hope is that academic and research institution staff will be able to teach their computers to read and process enormous amounts of scans, to confirm radiologists’ results, and potentially identify anything that may have been overlooked.
The NIH says in addition to being a “virtual radiology resident,” advanced computer technology has other potential benefits: For instance, it could identify slow changes occurring over the course of multiple chest x-rays that might otherwise be overlooked. The technology also would be useful in poor countries that lack radiologists. And in the future, the “resident” might be taught to read more complex images, such as CT and MRI.
The dataset, compiled from scans from > 30,000 patients, including many with advanced lung disease, was scrubbed of private information before release. The images are available via Box at https://nihcc.app.box.com/v/ChestXray-NIHCC.
Researchers around the world may be able to teach computers how to better detect and diagnose disease, thanks to > 100,000 chest x-ray images and corresponding data recently released by the NIH Clinical Center.
Reading and diagnosing chest x-rays requires careful observation, as well as knowledge of anatomy, physiology, and pathology. When that is combined with the need to consider all common thoracic diseases, it becomes hard to automate a consistent technique for reading images, the NIH says. With the free dataset, the hope is that academic and research institution staff will be able to teach their computers to read and process enormous amounts of scans, to confirm radiologists’ results, and potentially identify anything that may have been overlooked.
The NIH says in addition to being a “virtual radiology resident,” advanced computer technology has other potential benefits: For instance, it could identify slow changes occurring over the course of multiple chest x-rays that might otherwise be overlooked. The technology also would be useful in poor countries that lack radiologists. And in the future, the “resident” might be taught to read more complex images, such as CT and MRI.
The dataset, compiled from scans from > 30,000 patients, including many with advanced lung disease, was scrubbed of private information before release. The images are available via Box at https://nihcc.app.box.com/v/ChestXray-NIHCC.
Researchers around the world may be able to teach computers how to better detect and diagnose disease, thanks to > 100,000 chest x-ray images and corresponding data recently released by the NIH Clinical Center.
Reading and diagnosing chest x-rays requires careful observation, as well as knowledge of anatomy, physiology, and pathology. When that is combined with the need to consider all common thoracic diseases, it becomes hard to automate a consistent technique for reading images, the NIH says. With the free dataset, the hope is that academic and research institution staff will be able to teach their computers to read and process enormous amounts of scans, to confirm radiologists’ results, and potentially identify anything that may have been overlooked.
The NIH says in addition to being a “virtual radiology resident,” advanced computer technology has other potential benefits: For instance, it could identify slow changes occurring over the course of multiple chest x-rays that might otherwise be overlooked. The technology also would be useful in poor countries that lack radiologists. And in the future, the “resident” might be taught to read more complex images, such as CT and MRI.
The dataset, compiled from scans from > 30,000 patients, including many with advanced lung disease, was scrubbed of private information before release. The images are available via Box at https://nihcc.app.box.com/v/ChestXray-NIHCC.