I can see clearly now: image restoration via de-raining
We present a method for improving segmentation tasks on images affected by adherent rain drops and streaks. We introduce a novel stereo dataset recorded using a system that allows one lens to be affected by real water droplets while keeping the other lens clear. We train a denoising generator using...
| Main Authors: | , , |
|---|---|
| Format: | Conference item |
| Published: |
Institute of Electrical and Electronics Engineers
2019
|
| _version_ | 1797091267697442816 |
|---|---|
| author | Porav, H Bruls, T Newman, P |
| author_facet | Porav, H Bruls, T Newman, P |
| author_sort | Porav, H |
| collection | OXFORD |
| description | We present a method for improving segmentation tasks on images affected by adherent rain drops and streaks. We introduce a novel stereo dataset recorded using a system that allows one lens to be affected by real water droplets while keeping the other lens clear. We train a denoising generator using this dataset and show that it is effective at removing the effect of real water droplets, in the context of image reconstruction and road marking segmentation. To further test our de-noising approach, we describe a method of adding computer-generated adherent water droplets and streaks to any images, and use this technique as a proxy to demonstrate the effectiveness of our model in the context of general semantic segmentation. We benchmark our results using the CamVid road marking segmentation dataset, Cityscapes semantic segmentation datasets and our own realrain dataset, and show significant improvement on all tasks. |
| first_indexed | 2024-03-07T03:30:35Z |
| format | Conference item |
| id | oxford-uuid:ba94d2b1-8023-4183-8def-9bdb5abbafbf |
| institution | University of Oxford |
| last_indexed | 2024-03-07T03:30:35Z |
| publishDate | 2019 |
| publisher | Institute of Electrical and Electronics Engineers |
| record_format | dspace |
| spelling | oxford-uuid:ba94d2b1-8023-4183-8def-9bdb5abbafbf2022-03-27T05:10:51ZI can see clearly now: image restoration via de-rainingConference itemhttp://purl.org/coar/resource_type/c_5794uuid:ba94d2b1-8023-4183-8def-9bdb5abbafbfSymplectic Elements at OxfordInstitute of Electrical and Electronics Engineers2019Porav, HBruls, TNewman, PWe present a method for improving segmentation tasks on images affected by adherent rain drops and streaks. We introduce a novel stereo dataset recorded using a system that allows one lens to be affected by real water droplets while keeping the other lens clear. We train a denoising generator using this dataset and show that it is effective at removing the effect of real water droplets, in the context of image reconstruction and road marking segmentation. To further test our de-noising approach, we describe a method of adding computer-generated adherent water droplets and streaks to any images, and use this technique as a proxy to demonstrate the effectiveness of our model in the context of general semantic segmentation. We benchmark our results using the CamVid road marking segmentation dataset, Cityscapes semantic segmentation datasets and our own realrain dataset, and show significant improvement on all tasks. |
| spellingShingle | Porav, H Bruls, T Newman, P I can see clearly now: image restoration via de-raining |
| title | I can see clearly now: image restoration via de-raining |
| title_full | I can see clearly now: image restoration via de-raining |
| title_fullStr | I can see clearly now: image restoration via de-raining |
| title_full_unstemmed | I can see clearly now: image restoration via de-raining |
| title_short | I can see clearly now: image restoration via de-raining |
| title_sort | i can see clearly now image restoration via de raining |
| work_keys_str_mv | AT poravh icanseeclearlynowimagerestorationviaderaining AT brulst icanseeclearlynowimagerestorationviaderaining AT newmanp icanseeclearlynowimagerestorationviaderaining |