The paper says you used a C3D pretrained model but we can't find it anywhere about anomalydetectioncvpr2018

[kjsr7]

You should be using the Sports1m one.
First you need to extract the features using the C3D-V1.0 and then run the 'Save_C3DFeatures_32Segments.m' of this repository to save the features into 32 segments for each video. This technically explains how to extract the features.

[VVVVictorJ]

Hi,
I think you may be talk about the pretrained-model. https://www.dropbox.com/s/vr8ckp0pxgbldhs/conv3d_deepnetA_sport1m_iter_1900000?dl=0

[mourad1081]

Thank you both of you for your answers, they both help me.

[wzy1845]

Hi,
I think you may be talk about the pretrained-model. https://www.dropbox.com/s/vr8ckp0pxgbldhs/conv3d_deepnetA_sport1m_iter_1900000?dl=0

Hello, how to use this pre training model to extract features? Is there any code?thanks!