def _syntheticProc(dkf, dataset, mask, dataset_eval, mask_eval):
"""
Collect statistics on the synthetic dataset
"""
allmus, alllogcov = [], []
for s in range(10):
_, mus, logcov = DKF_evaluate.infer(dkf, dataset, mask)
allmus.append(np.copy(mus))
alllogcov.append(np.copy(logcov))
allmus_v, alllogcov_v = [], []
for s in range(10):
_, mus, logcov = DKF_evaluate.infer(dkf, dataset_eval, mask)
allmus_v.append(np.copy(mus))
alllogcov_v.append(np.copy(logcov))
mu_train = np.concatenate(allmus, axis=2).mean(2, keepdims=True)
cov_train = np.exp(np.concatenate(alllogcov, axis=2)).mean(2,
keepdims=True)
mu_valid = np.concatenate(allmus_v, axis=2).mean(2, keepdims=True)
cov_valid = np.exp(np.concatenate(alllogcov_v, axis=2)).mean(2,
keepdims=True)
#Extract the learned parameters w/in the generative model
learned_params = {}
for k in dkf.params_synthetic[dkf.params['dataset']]['params']:
learned_params[k] = dkf.tWeights[k + '_W'].get_value()
return mu_train, cov_train, mu_valid, cov_valid, learned_params
def knn(query, samples, neighbor_by_class=3):
"""Infer a query with K-nearsest neighbors strategy
Args:
query (dict): { 'img': <path_to_img>, 'cls': <img class>, 'hist' <img histogram> }
samples (list): list of { 'img': <path_to_img>, 'cls': <img class>, 'hist' <img histogram> }
neighbor_by_class (int, optional): nb nearsest neighbors by class. Defaults to 3.
Returns:
str: selected class
"""
_, results = infer(query, samples)
# Group by class
res = sorted(results, key=lambda x: x['cls'])
grouped_by_class = {
k: list(sorted(map(lambda x: x['dis'], g)))[:neighbor_by_class]
for k, g in groupby(res, key=lambda x: x['cls'])
}
# Calc average by class
grouped_by_class = {k: mean(v) for k, v in grouped_by_class.items()}
# Get max
selected_class = sorted(grouped_by_class.items(),
key=lambda x: x[1])[:1][0][0]
return selected_class
def get_results():
if request.method == 'POST':
q_img = request.files['image']
print(q_img)
obj = Embeddings()
q_hist = obj.make_hist(q_img)
results = infer(q_hist, depth=depth, d_type=d_type)
return render_template('results.html', results=results)
def _syntheticProc(dkf, dataset, dataset_eval):
"""
Collect statistics on the synthetic dataset
"""
allmus, alllogcov = [], []
for s in range(100):
_,mus, logcov = DKF_evaluate.infer(dkf,dataset)
allmus.append(np.copy(mus))
alllogcov.append(np.copy(logcov))
allmus_v, alllogcov_v = [], []
for s in range(100):
_,mus, logcov = DKF_evaluate.infer(dkf,np.copy(dataset_eval))
allmus_v.append(np.copy(mus))
alllogcov_v.append(np.copy(logcov))
mu_train = np.concatenate(allmus,axis=2).mean(2,keepdims=True)
cov_train= np.exp(np.concatenate(alllogcov,axis=2)).mean(2,keepdims=True)
mu_valid = np.concatenate(allmus_v,axis=2).mean(2,keepdims=True)
cov_valid= np.exp(np.concatenate(alllogcov_v,axis=2)).mean(2,keepdims=True)
return mu_train, cov_train, mu_valid, cov_valid
# print(result)
# # retrieve by gabor
# method = Gabor()
# samples = method.make_samples(db)
# query = samples[query_idx]
# _, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# print(result)
# # retrieve by HOG
# method = HOG()
# samples = method.make_samples(db)
# query = samples[query_idx]
# _, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# print(result)
# # retrieve by VGG
# method = VGGNetFeat()
# samples = method.make_samples(db)
# query = samples[query_idx]
# _, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# print(result)
# retrieve by resnet
method = ResNetFeat()
samples = method.make_samples(db)
query_idx = 1569
query = samples[query_idx]
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
print(result)
from evaluate import infer
from create_hist import Embeddings
import cv2
import time
depth = 15
d_type = 'd1'
q_img = "database/images/3159.jpg"
if __name__ == '__main__':
obj = Embeddings()
q_hist = obj.make_hist(q_img)
i_img = cv2.imread(q_img)
cv2.imshow('Query', i_img)
results = infer(q_hist, depth=depth, d_type=d_type)
for index, result in enumerate(results):
print(index, result['img'], result['dis'])
r = cv2.imread(result['img'])
cv2.imshow("Results", r)
cv2.waitKey(0)
samples.append({'img': d_img, 'cls': d_cls, 'hist': d_hist})
cPickle.dump(samples,
open(os.path.join(cache_dir, sample_cache), "wb"))
return samples
if __name__ == "__main__":
d = Daisy()
# Create my samples
db = Database("database\\train")
print("Train databse created.")
samples = d.make_samples(db, sample_name="train")
print("Train samples created.")
test = Database("database\dev")
print("Test databse created.")
sample_test = d.make_samples(test, sample_name="dev")
print("Test samples created.")
# Find class for each image of my test DB and verify the result
nb_good_classification = 0
for img_test in sample_test:
_, resultes = infer(img_test, samples)
real_cls = KNN(resultes, db.get_class())
nb_good_classification += get_cls(img_test['cls']) == get_cls(real_cls)
print("\n{}/{}".format(nb_good_classification, len(sample_test)))
def test(db, query_idx):
results = {}
# retrieve by color
method = Color()
samples = method.make_samples(db)
query = samples[query_idx]
# print(samples)
img = scipy.misc.imread(query['img'])
# print(query)
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# results.append(result[0]['cls'])
inc(results, result[0]['cls'])
# # retrieve by daisy
# method = Daisy()
# samples = method.make_samples(db)
# query = samples[query_idx]
# _, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# # results.append(result[0]['cls'])
# inc(results, result[0]['cls'])
# # # retrieve by edge
method = Edge()
samples = method.make_samples(db)
query = samples[query_idx]
# print(samples)
query = samples[query_idx]
img = scipy.misc.imread(query['img'])
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# results.append(result[0]['cls'])
inc(results, result[0]['cls'])
# # # retrieve by gabor
# # method = Gabor()
# # samples = method.make_samples(db)
# # query = samples[query_idx]
# # _, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# # print(result)
# # inc(results, result[0]['cls'])
# # retrieve by HOG
method = HOG()
samples = method.make_samples(db)
query = samples[query_idx]
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# results.append(result[0]['cls'])
inc(results, result[0]['cls'])
# # retrieve by VGG
method = VGGNetFeat()
samples = method.make_samples(db)
query = samples[query_idx]
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# results.append(result[0]['cls'])
inc(results, result[0]['cls'])
# # retrieve by resnet
method = ResNetFeat()
samples = method.make_samples(db)
query = samples[query_idx]
_, result = infer(query, samples=samples, depth=depth, d_type=d_type)
# results.append(result[0]['cls'])
inc(results, result[0]['cls'])
import os
from PIL import Image
print(results)
finalresult = max(results.items(), key=operator.itemgetter(1))[0]
#string=".../database/"+finalresult+"/"
string = "./database/" + finalresult + "/"
print(string)
a = 1
for file in os.listdir(string):
a += 1
tempimg = Image.open(string + file)
tempimg.show()
print(string + file)
if (a == 10):
break
print("Final result is: ", finalresult)
scipy.misc.imshow(img)
