|
|
import h5py
|
|
|
import torch as t
|
|
|
import numpy as np
|
|
|
from segment import cpd_auto
|
|
|
|
|
|
max_segment_num = 20
|
|
|
max_frame_num = 200
|
|
|
|
|
|
for kind in ["clip_ViT-B16_norm_avg"]:
|
|
|
print(kind)
|
|
|
for video_id in ["V1","V2","V3","V4"]:
|
|
|
f=h5py.File('./features/'+video_id+'_'+kind+'.h5','r')
|
|
|
feature=f['feature'][()]
|
|
|
frame_num=feature.shape[0]
|
|
|
print(frame_num)
|
|
|
|
|
|
K=feature
|
|
|
K=np.dot(K,K.T)
|
|
|
|
|
|
cps,_=cpd_auto(K,max_segment_num-1,1,desc_rate=1,verbose=False,lmax=max_frame_num-1)
|
|
|
seg_num=len(cps)+1
|
|
|
|
|
|
assert seg_num<=max_segment_num
|
|
|
|
|
|
seg_points=cps
|
|
|
seg_points=np.insert(seg_points,0,0)
|
|
|
seg_points=np.append(seg_points,frame_num)
|
|
|
|
|
|
segments=[]
|
|
|
for i in range(seg_num):
|
|
|
segments.append(np.arange(seg_points[i],seg_points[i+1],1,dtype=np.int32))
|
|
|
|
|
|
assert len(segments)<=max_segment_num
|
|
|
|
|
|
for seg in segments:
|
|
|
assert len(seg)<=max_frame_num
|
|
|
|
|
|
seg_len=np.zeros((max_segment_num),dtype=np.int32)
|
|
|
for index,seg in enumerate(segments):
|
|
|
seg_len[index]=len(seg)
|
|
|
|
|
|
|
|
|
|
|
|
for seg in segments:
|
|
|
for frame in seg:
|
|
|
assert frame<frame_num
|
|
|
|
|
|
if kind=="clip_ViT-B16_norm_avg":
|
|
|
feature_dim=512
|
|
|
elif kind=="clip_ViT-B32_norm_avg":
|
|
|
feature_dim=512
|
|
|
elif kind=="clip_ViT-L14_norm_avg":
|
|
|
feature_dim=768
|
|
|
else:
|
|
|
feature_dim=768
|
|
|
|
|
|
features=t.zeros((max_segment_num, max_frame_num, feature_dim))
|
|
|
for seg_index,seg in enumerate(segments):
|
|
|
for frame_index,frame in enumerate(seg):
|
|
|
features[seg_index,frame_index]=t.tensor(feature[frame])
|
|
|
|
|
|
|
|
|
|
|
|
f=h5py.File('./processed/'+video_id+'_'+kind+'.h5','w')
|
|
|
f.create_dataset('features', data=features)
|
|
|
f.create_dataset('seg_len', data=seg_len)
|
|
|
|
|
|
f.close()
|
|
|
|
