def task2(self, k=10):
QS2 = HistDataset(self.QSD1_W3,
method="combo",
masking=False,
bbox=True,
multires=4,
denoise=True,
texture="LBP")
GT = get_groundtruth("datasets/qsd1_w3/gt_corresps.pkl")
print(f"Computing normalized histograms for {self.DDBB}")
DB = list(
tqdm(
HistDataset(self.DDBB,
masking=False,
method="combo",
multires=4,
texture="LBP")))
print("Analyzing QS2")
find_img_corresp(QS2, GT, DB, k)
def test_poor_read(self):
d = HistDataset(cfg.IMAGE_FOLDER, cfg.LABEL_FOLDER, cfg.LABEL_2_ID)
loader = DataLoader(d,
batch_size=1,
num_workers=4,
collate_fn=lambda x: x[0])
sli = itertools.islice(loader, 11000, 12000)
for sample in sli:
img = sample.image
def task5(self):
# Get text box pkl
QS = [
text_removal.getpoints2(im)
for im in text_removal.text_remover(self.QSD1_W2)
]
boundingxys = [[element.boundingxy] for element in QS]
with open("QSD1/text_boxes.pkl", "wb") as f:
pickle.dump(boundingxys, f)
# Get text box pngs TODO bbox
QS1 = HistDataset("datasets/qsd1_w2", bbox=True, multires=2)
predicted_masks = [QS1.get_mask(idx) for idx in range(len(QS1))]
for i, img in enumerate(predicted_masks):
filename = "QSD1/boxes/" + f"{i:05d}" + ".png"
cv2.imwrite(filename, img)
gt = np.asarray(
get_groundtruth("datasets/qsd1_w2/text_boxes.pkl")).squeeze()
mean_IoU = get_mean_IoU(gt, boundingxys)
print(f"Mean IoU: {mean_IoU}")
def test_run(cfg):
dataset = HistDataset(cfg['CLASS_TYPE'],
get_path(cfg['TEST_FOLDER']),
get_path(cfg['LABEL_FOLDER']),
split=[0, 0, 100],
label2id=cfg['LABEL_2_ID'],
randomize=cfg['INIT_RANDOMIZE'])
print("Prepping Dataset ...")
test_loader = DataLoader(dataset.test,
batch_size=cfg['BATCH_SIZE'],
num_workers=cfg['NUM_WORKERS'],
collate_fn=batchify)
model = torch.load(cfg['MODEL_SAVE_FILE'])
test_a_epoch("ultimate_test", test_loader, model, cfg['ULTI_RESULT_FILE'],
cfg, 0)
def task6(self, k=10):
QS = [ # noqa
hists for hists in tqdm(
MultiHistDataset(self.QSD2_W3,
masking=True,
bbox=True,
multires=4,
method="color",
texture="LBP",
denoise=True))
]
GT = get_pickle("datasets/qsd2_w3/gt_corresps.pkl")
DB = list(
tqdm(
HistDataset(self.DDBB,
masking=False,
multires=4,
method="color",
texture="LBP"))) # noqa
tops = find_multi_img_corresp_keep(QS, DB, k)
dump_pickle("result_qst2.pkl", tops)
mapAtK = metrics.mapk(GT, tops, k)
print("Map@k is " + str(mapAtK))
exit()
with open("outputs/resutls.pkl", "wb") as f:
pickle.dump(tops, f)
print(tops)
# Generate pngs
QS1 = Dataset(self.QSD2_W2, masking=True, bbox=True)
for i in range(len(QS1)):
im = QS1.get_mask(i)
cv2.imwrite("outputs/" + str(i) + ".png", im)
text_boxes = [BBox().get_bbox_cords(QS1[i]) for i in range(len(QS1))]
with open("outputs/text_boxes.pkl", "wb") as f:
pickle.dump(text_boxes, f)
def test_label_conversion(self):
d = HistDataset(cfg.IMAGE_FOLDER, cfg.LABEL_FOLDER)
loader = DataLoader(d,
batch_size=1,
num_workers=4,
collate_fn=lambda x: x[0])
def test_collection(self):
d = HistDataset().collection(cfg.IMAGE_FOLDER, cfg.LABEL_FOLDER)
self.__test_folder(d)
for s in d:
self.__test_image_size(s)