def launch(method, neighbor_by_class: int = 3, store: bool = False):
"""Launch a classification algorithm. Train on a first dataset, and validate it with a second (and print result)
Args:
method (object): Class object that have a make_samples function
neighbor_by_class (int, optional): Nb nearsest neighbors by class. Defaults to 3.
store (bool, optional): True if you want to store the alg result in a folder. Default to False.
"""
db_train = Database(DB_TRAIN)
db_validation = Database(DB_VALIDATION)
train_samples = method.make_samples(db_train)
validation_samples = method.make_samples(db_validation)
if store:
# Delete previous result
for classe in db_train.get_class():
shutil.rmtree(os.path.join(RESULT_RETRIEVAL, classe),
ignore_errors=True)
# and re-create directory
for classe in db_train.get_class():
os.mkdir(os.path.join(RESULT_RETRIEVAL, classe))
final = {cl: [0, 0] for cl in sorted(db_train.get_class())}
for img in validation_samples:
selected_class = knn(img,
train_samples,
neighbor_by_class=neighbor_by_class)
# Compare selected class with real class
if img['cls'] == selected_class:
final[img['cls']][0] += 1
final[img['cls']][1] += 1
# Store image class result if asked
if store:
# Copy image to result dir
i = os.path.basename(img['img'])
shutil.copy(img['img'],
os.path.join(RESULT_RETRIEVAL, selected_class, i))
# Beautiful print
print("\n========= Result =========")
col_width = max(len(cl) for cl, v in final.items()) + 2
for cl, v in final.items():
print(
f"{cl.ljust(col_width)}: {v[0]}/{v[1]}\t{round(v[0]*100/v[1], 2)} %"
)
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)))