File size: 1,654 Bytes
32b542e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 |
import os
import numpy as np
from .func_feats import boxes_to_locfeats
import pdb
def read_lines(path):
with open(path, 'r') as fid:
lines = [line.strip() for line in fid]
return lines
def read_lines_set(path):
lines = read_lines(path)
lines = set(lines)
return lines
# "features", "cls_prob", "boxes", "image_h", "image_w"
def read_np(path, preload=None):
if preload:
content = preload[path]
else:
content = np.load(path, allow_pickle=True)
if isinstance(content, np.ndarray):
return { "features": content }
keys = content.keys()
if len(keys) == 1:
return { "features": content[list(keys)[0]] }
return content
def read_np_bbox(path, max_feat_num, use_global_v=True, preload=None):
content = read_np(path, preload)
features = content['features'][0:max_feat_num - 1]
boxes = content['boxes'][0:max_feat_num - 1]
image_h = content['image_h'][0]
image_w = content['image_w'][0]
num_boxes = len(boxes)
if use_global_v:
g_feat = np.sum(features, axis=0) / num_boxes
features = np.concatenate([np.expand_dims(g_feat, axis=0), features], axis=0)
image_locations = boxes_to_locfeats(boxes, image_w, image_h)
if use_global_v:
g_location = np.array([0, 0, 1, 1, 1])
image_locations = np.concatenate([np.expand_dims(g_location, axis=0), image_locations], axis=0)
return features, image_locations
def load_vocab(path):
if len(path) == 0:
return None
vocab = ['.']
with open(path, 'r') as fid:
for line in fid:
vocab.append(line.strip())
return vocab |