SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2202.04901 · ECCV · 2022

FILM: Frame Interpolation for Large Motion

Deqing Sun, Brian Curless, Janne Kontkanen, Eric Tabellion, Fitsum Reda, Caroline Pantofaru

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 20 functions out of this paper's own repositories and ran 12 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
google-research/frame-interpolation canonical 2 of 6
dajes/frame-interpolation-pytorch — 10 of 14
FunctionStatusWhere it lives
Conv2d Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("33031ed5f24825b1")
FeatureExtractor Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("9ef67e49992f642f")
FlowEstimator Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("a7349eca770ecdf9")
Fusion Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("5df4eeacc27fbe49")
PyramidFlowEstimator Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("646c00a2c3ce3ec0")
SubTreeExtractor Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("c47da384c303149c")
build_image_pyramid Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("5e996c1e85b97cc1")
concatenate_pyramids Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("e5a70acb4070eb0f")
flow_pyramid_synthesis Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("6ea74e810cda7a86")
get_channels_at_level Ran dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("f8ee6422386d84b6")
l1_loss Ran google-research/frame-interpolation/losses/losses.py
code served (permissive licence) · get_code("36bbb37e74c4a4dc")
patches_to_image Ran google-research/frame-interpolation/eval/interpolator.py
code served (permissive licence) · get_code("41eb7c711cf87551")
Interpolator Not yet run dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("e48dec25e2c77c29")
generate_image_triplet_example Not yet run google-research/frame-interpolation/datasets/util.py
code served (permissive licence) · get_code("62b36be97a57cd56")
image_to_patches Not yet run google-research/frame-interpolation/eval/interpolator.py
code served (permissive licence) · get_code("81e6e74325f0e8ca")
multiply_pyramid Not yet run dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("12d32fa9bc246c8f")
pyramid_warp Not yet run dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("6c403027f71d5c0e")
style_loss Not yet run google-research/frame-interpolation/losses/vgg19_loss.py
code served (permissive licence) · get_code("c4b4c0fc22dc8ce1")
vgg_loss Not yet run google-research/frame-interpolation/losses/vgg19_loss.py
code served (permissive licence) · get_code("56ee3540e67f1685")
warp Not yet run dajes/frame-interpolation-pytorch/interpolator.py
code served (permissive licence) · get_code("b6d4f892f64d3fe9")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We present a frame interpolation algorithm that synthesizes an engaging slow-motion video from near-duplicate photos which often exhibit large scene motion. Near-duplicates interpolation is an interesting new application, but large motion poses challenges to existing methods. To address this issue, we adapt a feature extractor that shares weights across the scales, and present a "scale-agnostic" motion estimator. It relies on the intuition that large motion at finer scales should be similar to small motion at coarser scales, which boosts the number of available pixels for large motion supervision. To inpaint wide disocclusions caused by large motion and synthesize crisp frames, we propose to optimize our network with the Gram matrix loss that measures the correlation difference between features. To simplify the training process, we further propose a unified single-network approach that removes the reliance on additional optical-flow or depth network and is trainable from frame triplets alone. Our approach outperforms state-of-the-art methods on the Xiph large motion benchmark while performing favorably on Vimeo-90K, Middlebury and UCF101. Source codes and pre-trained models are available at https://film-net.github.io.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2202.04901")
get_code_for_paper("2202.04901")
have("2202.04901")

Connect an agent — have() is free.