{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/uncertainty-measures-and-prediction-quality","title":"Uncertainty Measures and Prediction Quality Rating for the Semantic Segmentation of Nested Multi Resolution Street Scene Images","arxiv_id":"1904.04516","date":"2019-04-09","proceeding":null,"authors":["Matthias Rottmann","Marius Schubert"],"abstract":"In the semantic segmentation of street scenes the reliability of the\nprediction and therefore uncertainty measures are of highest interest. We\npresent a method that generates for each input image a hierarchy of nested\ncrops around the image center and presents these, all re-scaled to the same\nsize, to a neural network for semantic segmentation. The resulting softmax\noutputs are then post processed such that we can investigate mean and variance\nover all image crops as well as mean and variance of uncertainty heat maps\nobtained from pixel-wise uncertainty measures, like the entropy, applied to\neach crop's softmax output. In our tests, we use the publicly available\nDeepLabv3+ MobilenetV2 network (trained on the Cityscapes dataset) and\ndemonstrate that the incorporation of crops improves the quality of the\nprediction and that we obtain more reliable uncertainty measures. These are\nthen aggregated over predicted segments for either classifying between IoU=0\nand IoU>0 (meta classification) or predicting the IoU via linear regression\n(meta regression). The latter yields reliable performance estimates for\nsegmentation networks, in particular useful in the absence of ground truth. For\nthe task of meta classification we obtain a classification accuracy of\n$81.93\\%$ and an AUROC of $89.89\\%$. For meta regression we obtain an $R^2$\nvalue of $84.77\\%$. These results yield significant improvements compared to\nother approaches.","url_abs":"http://arxiv.org/abs/1904.04516v1","url_pdf":"http://arxiv.org/pdf/1904.04516v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"uncertainty-measures-and-prediction-quality","repo_url":"https://github.com/mrottmann/MetaSeg","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"depthwise-convolution","method_name":"Depthwise Convolution"},{"method_slug":"depthwise-separable-convolution","method_name":"Depthwise Separable Convolution"},{"method_slug":"inverted-residual-block","method_name":"Inverted Residual Block"},{"method_slug":"pointwise-convolution","method_name":"Pointwise Convolution"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}