Budget-aware Semi-Supervised Semantic and Instance Segmentation - PowerPoint PPT Presentation

Budget-aware Semi-Supervised Semantic and Instance Segmentation Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto Women In Computer Vision - CVPR 2019

Motivation Semantic segmentation Instance segmentation Pixel-level annotations are expensive! Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Motivation Semantic segmentation Instance segmentation Solution: Weakly and Semi-supervised methods! Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Motivation Semantic segmentation Instance segmentation Solution: Weakly and Semi-supervised methods! Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Contributions 1. We unify the segmentation benchmarks regardless the training setting and the supervision signals comparing them in terms of the total annotation cost . Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Contributions 1. We unify the segmentation benchmarks regardless the training setting and the supervision signals comparing them in terms of the total annotation cost . 2. We experiment with a semi-supervised pipeline and test it in Pascal VOC for semantic and instance segmentation, outperforming previous works at low annotated budgets. Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Contributions 1. We unify the segmentation benchmarks regardless the training setting and the supervision signals comparing them in terms of the total annotation cost . 2. We experiment with a semi-supervised pipeline and test it in Pascal VOC for semantic and instance segmentation, outperforming previous works at low annotated budgets. 3. We show that for low annotation budgets, it’s more convenient having fewer but stronger-labeled data over having larger weakly-annotated sets . Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Annotation Semi-Supervised Pipeline Semi-Supervised Pipeline network Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Annotation Semi-Supervised Pipeline Semi-Supervised Pipeline network Segmentation network Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Experimental validation for Pascal VOC Semantic segmentation Segmentation quality We used DeepLab v3+ for both the annotation and segmentation network Annotation cost Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Experimental validation for Pascal VOC Instance segmentation Segmentation quality We used RSIS for both the annotation and segmentation network Annotation cost Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Comparison to other works Semantic segmentation Instance segmentation Segmentation quality Annotation cost Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Visualization for semantic segmentation Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Visualization for instance segmentation Budget-aware Semi-Supervised Semantic and Instance Segmentation : Miriam Bellver, Amaia Salvador, Jordi Torres, Xavier Giro-i-Nieto

Thanks for your attention!