"""Copyright @ Xianming Liu, University of Illinois at Urbana-Champaign
MultiLabelLayer, which deals with the multiple labels input,
Outputs include both data and label numpy arrays,
For the labels, the output will be a probabilistic array or binary array
"""
import numpy as np
from BasePythonDataLayer import BasePythonDataLayer
from utils.SampleIO import extract_sample
from utils.util import parse_label
__author__ = ['Xianming Liu(liuxianming@gmail.com)']
class MultiLabelLayer(BasePythonDataLayer):
"""Implementation of Data Layer for MultiLabel raw data
"""
def setup(self, bottom, top):
super(MultiLabelLayer, self).setup(bottom, top)
self._multilabel = self._layer_params.get('multilabel', False)
if not self._multilabel:
self._label_dim = 1
else:
self._label_dim = self._layer_params.get('label dim', None)
if not self._label_dim:
# try to estimate the dimension of labels
self.calculate_label_dim()
self._cur = 0
def calculate_label_dim(self):
"""Calculate the dimension of labels
by calculating the lenth of label set
"""
all_labels = []
for label_str in self._label:
label = parse_label(label_str)
all_labels += label
all_labels = set(all_labels)
self._label_dim = len(all_labels)
def get_a_datum(self):
if self._compressed:
datum = extract_sample(
self._data[self._cur], self._mean, self._resize)
else:
datum = self._data[self._cur]
# start parsing labels
label_elems = parse_label(self._label[self._cur])
label = np.zeros(self._label_dim)
if not self._multilabel:
label[0] = label_elems[0]
else:
for i in label_elems:
label[i] = 1
self._cur = (self._cur + 1) % self._sample_count
return datum, label
def get_next_minibatch(self):
data = []
labels = []
for i in range(self._batch_size):
datum, label = self.get_a_datum()
data.append(datum)
labels.append(label)
batch = [
np.array(data),
np.array(labels).reshape(self._batch_size, self._label_dim, 1, 1)
]
return batch
