idx = path.split(os.path.sep)[-1].split(".")[0]
ids.append(int(idx))
# preprocess image
image = cv2.imread(path)
for p in [aap, iap]: # maintain AR and resize to 224x224
image = p.preprocess(image)
# Special for ImageNet dataset => substracting mean RGB pixel intensity
image = imagenet_utils.preprocess_input(image) # (224, 224, 3)
batchImages.append(image)
pass
if useTTA == "True":
for image in batchImages:
crops = cp1.preprocess(image)
# predict over 10 crops
crops_probs = model.predict(crops)
# predict probs of dogs
pred = crops_probs.mean(axis=0)[1]
predictions.append(pred)
pass
else:
# loop over batchImages, resize to (227, 227)
for i in range(len(batchImages)):
batchImages[i] = cv2.resize(batchImages[i], (224, 224))
batchImages = np.array(batchImages) # (224, 224, 3)
for i in tqdm(range(0, N, B)):
batchPaths = imagePaths[i:i + B]
batchImages = []
for path in batchPaths:
image = cv2.imread(path)
image = aap.preprocess(
image) # maintain AR and resize to 256 x 256
image = iap.preprocess(image)
# Special for ImageNet dataset => substracting mean RGB pixel intensity
image = imagenet_utils.preprocess_input(image) # (256, 256, 3)
batchImages.append(image)
pass
if useTTA == "True":
for image in batchImages:
crops = cp1.preprocess(image)
# predict over 10 crops
crops_probs = model.predict(crops)
#pred = crops_probs.mean(axis=0) # (1. 2)
#predictions.append(np.argmax(pred))
# predict probs of dogs
pred = crops_probs.mean(axis=0)[1]
predictions.append(pred)
pass
else:
# loop over batchImages, resize to (227, 227)
for i in range(len(batchImages)):