Store and Fetch Immediately: Everything Is All You Need for Space-Time Video Super-resolution

نویسندگان

چکیده

Existing space-time video super-resolution (ST-VSR) methods fail to achieve high-quality reconstruction since they fully explore the spatial-temporal correlations, long-range components in particular. Although recurrent structure for ST-VSR adopts bidirectional propagation aggregate information from entire video, collecting temporal between past and future via one-stage representations inevitably loses relations. To alleviate limitation, this paper proposes an immediate storeand-fetch network promote correlation learning, where stored can be refetched help representation of current frame. Specifically, proposed consists two modules: a backward module (BRM) forward (FRM). The former first performs inference past, while storing (SR) each Following that, latter super-resolve all frames, SR Since FRM inherits BRM, therefore, spatial sequence is immediately fetched, which allows drastic improvement ST-VSR. Extensive experiments both on space (S-VSR) as well time (T-VSR) have demonstrated effectiveness our method over other state-of-the-art public datasets. Code available https://github.com/hhhhhumengshun/SFI-STVR

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

All You Need Is Mentorship

I find it humbling to confess that most of the truly original ideas that have driven my research group’s agenda over four decades of time have come, not from my own brain, but instead from the minds of my trainees, both graduate students and post-docs. This on its own might explain why I, rather selfishly, have given them long leashes, allowing them to strike out on their own and craft their ow...

متن کامل

All You Need Is Compassion

The paper presents a new deductive rule for verifying response properties under the assumption of compassion (strong fairness) requirements. It improves on previous rules in that the premises of the new rule are all first order. We prove that the rule is sound, and present a constructive completeness proof for the case of finite-state systems. For the general case, we present a sketch of a rela...

متن کامل

CNN Is All You Need

CNNs have been successfully used in audio, image and text classification, analysis and generation [12,17,18], whereas the RNNs with LSTM cells [5,6] have been widely adopted for solving sequence transduction problems such as language modeling and machine translation [19,3,5]. The RNN models typically align the element positions of the input and output sequences to steps in computation time for ...

متن کامل

Attention is All you Need

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. E...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Proceedings of the ... AAAI Conference on Artificial Intelligence

سال: 2023

ISSN: ['2159-5399', '2374-3468']

DOI: https://doi.org/10.1609/aaai.v37i1.25165