The performance of video frame interpolation is inherently correlated with the ability to
handle motion in the input scene. Even though previous works recognize the utility of
asynchronous event information for this task, they ignore the fact that motion may or may not
result in blur in the input video to be interpolated, depending on the length of the exposure
time of the frames and the speed of the motion, and assume either that the input video is
sharp, restricting themselves to frame interpolation, or that it is blurry, including an explicit …

Fig. 7 shows the detailed architecture of the proposed Event Recurrent (EVR) block. The
feature maps from the previous blocks are concatenated with the previous state, and then
sent to the residual blocks. In the end, after the convolutional layer and the activation layer,
the new state is propagated to the next recurrent iteration. The feature maps are down-
sampled to half of the original size, and sent to the next block.