Predicting scoliosis in DXA scans using intermediate representations
We describe a method to automatically predict scoliosis in Dual-energy X-ray Absorptiometry (DXA) scans. We also show that in- termediate representations, which in our case are segments of body parts, help improve performance. Hence, we propose a two step process for pre- diction: (i) we learn to se...
| Главные авторы: | , , , |
|---|---|
| Формат: | Conference item |
| Опубликовано: |
Springer
2019
|
| _version_ | 1826303456964509696 |
|---|---|
| author | Jamaludin, A Kadir, T Clark, E Zisserman, A |
| author_facet | Jamaludin, A Kadir, T Clark, E Zisserman, A |
| author_sort | Jamaludin, A |
| collection | OXFORD |
| description | We describe a method to automatically predict scoliosis in Dual-energy X-ray Absorptiometry (DXA) scans. We also show that in- termediate representations, which in our case are segments of body parts, help improve performance. Hence, we propose a two step process for pre- diction: (i) we learn to segment body parts via a segmentation Convolutional Neural Network (CNN), which we show outperforms the noisy labels it was trained on, and (ii) we predict with a classification CNN that uses as input both the raw DXA scan and also the intermediate representation, i.e. the segmented body parts. We demonstrate that this two step process can predict scoliosis with high accuracy, and can also localize the spinal curves (i.e. geometry) without additional supervision. Furthermore, we also propose a soft score of scoliosis based on the classification CNN which correlates to the severity of scoliosis. |
| first_indexed | 2024-03-07T06:02:58Z |
| format | Conference item |
| id | oxford-uuid:ecdb963c-4050-4de2-9d04-489413d4cfa0 |
| institution | University of Oxford |
| last_indexed | 2024-03-07T06:02:58Z |
| publishDate | 2019 |
| publisher | Springer |
| record_format | dspace |
| spelling | oxford-uuid:ecdb963c-4050-4de2-9d04-489413d4cfa02022-03-27T11:20:36ZPredicting scoliosis in DXA scans using intermediate representationsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:ecdb963c-4050-4de2-9d04-489413d4cfa0Symplectic Elements at OxfordSpringer2019Jamaludin, AKadir, TClark, EZisserman, AWe describe a method to automatically predict scoliosis in Dual-energy X-ray Absorptiometry (DXA) scans. We also show that in- termediate representations, which in our case are segments of body parts, help improve performance. Hence, we propose a two step process for pre- diction: (i) we learn to segment body parts via a segmentation Convolutional Neural Network (CNN), which we show outperforms the noisy labels it was trained on, and (ii) we predict with a classification CNN that uses as input both the raw DXA scan and also the intermediate representation, i.e. the segmented body parts. We demonstrate that this two step process can predict scoliosis with high accuracy, and can also localize the spinal curves (i.e. geometry) without additional supervision. Furthermore, we also propose a soft score of scoliosis based on the classification CNN which correlates to the severity of scoliosis. |
| spellingShingle | Jamaludin, A Kadir, T Clark, E Zisserman, A Predicting scoliosis in DXA scans using intermediate representations |
| title | Predicting scoliosis in DXA scans using intermediate representations |
| title_full | Predicting scoliosis in DXA scans using intermediate representations |
| title_fullStr | Predicting scoliosis in DXA scans using intermediate representations |
| title_full_unstemmed | Predicting scoliosis in DXA scans using intermediate representations |
| title_short | Predicting scoliosis in DXA scans using intermediate representations |
| title_sort | predicting scoliosis in dxa scans using intermediate representations |
| work_keys_str_mv | AT jamaludina predictingscoliosisindxascansusingintermediaterepresentations AT kadirt predictingscoliosisindxascansusingintermediaterepresentations AT clarke predictingscoliosisindxascansusingintermediaterepresentations AT zissermana predictingscoliosisindxascansusingintermediaterepresentations |