Novel view synthesis via depth-guided skip connections

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference contributionScientificvertaisarvioitu


We introduce a principled approach for synthesizing new views of a scene given a single source image. Previous methods for novel view synthesis can be divided into image-based rendering methods (e.g., flow prediction) or pixel generation methods. Flow predictions enable the target view to re-use pixels directly, but can easily lead to distorted results. Directly regressing pixels can produce structurally consistent results but generally suffer from the lack of low-level details. In this paper, we utilize an encoder-decoder architecture to regress pixels of a target view. In order to maintain details, we couple the decoder aligned feature maps with skip connections, where the alignment is guided by predicted depth map of the target view. Our experimental results show that our method does not suffer from distortions and successfully preserves texture details with aligned skip connections.

OtsikkoProceedings - 2021 IEEE Winter Conference on Applications of Computer Vision, WACV 2021
ISBN (elektroninen)9780738142661
DOI - pysyväislinkit
TilaJulkaistu - tammikuuta 2021
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisuussa
TapahtumaIEEE Winter Conference on Applications of Computer Vision - Virtual, Online, Yhdysvallat
Kesto: 5 tammikuuta 20219 tammikuuta 2021


ConferenceIEEE Winter Conference on Applications of Computer Vision
KaupunkiVirtual, Online


Sukella tutkimusaiheisiin 'Novel view synthesis via depth-guided skip connections'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä