def pascal_voc_overlay(the_image):
segment_image = semantic_segmentation()
segment_image.load_pascalvoc_model(
"model/deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
segment_image.segmentAsPascalvoc(the_image,
output_image_name='res' + '/' +
f'{the_image}',
overlay=True)
st.image(f'res/{the_image}', caption='Output Image')
def sem_seg():
capture = cv2.VideoCapture(0)
segment_video = semantic_segmentation()
segment_video.load_ade20k_model(
"deeplabv3_xception65_ade20k.h5") #loading_the_datamodel
segment_video.process_camera_ade20k(capture,
overlay=True,
frames_per_second=10,
output_video_name="sem_seg_out.mp4",
show_frames=True,
frame_name="frame")
def deep_segmentation(self, pic):
path = '/Users/regevazran/Desktop/technion/semester i/project c/temp pic/'
result = cv2.imwrite(path + 'temp_pic.jpg', pic)
from pixellib.semantic import semantic_segmentation
segment_image = semantic_segmentation()
segment_image.load_pascalvoc_model(
path + "deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
segment_image.segmentAsPascalvoc(path + 'temp_pic.jpg',
output_image_name=path +
"image_new.jpg")
return
def post(self):
if request.json:
image = request.json['image']
image_string = base64.b64decode(image)
image_data = BytesIO(image_string)
img = Image.open(image_data)
img.save(INPUT_IMAGE)
semantic_segment_image = semantic_segmentation()
semantic_segment_image.load_pascalvoc_model(SEMANTIC_MODEL)
semantic_segment_image.segmentAsPascalvoc(
INPUT_IMAGE, output_image_name=OUTPUT_IMAGE)
with open(OUTPUT_IMAGE, "rb") as img_file:
my_string = base64.b64encode(img_file.read())
final_base64_image_string = my_string.decode('utf-8')
return {"output_image": final_base64_image_string}
def prediction(img_file):
# instantiating the semantic segmentation class
segment_image = semantic_segmentation()
# loading the model deeplabv3+ trained on pascal voc dataset.
segment_image.load_pascalvoc_model(DATA_URL)
# performing the segmentation on the input image
segment_image.segmentAsPascalvoc(img_file, output_image_name = "output_images/out.jpg")
out = plt.imread("output_images/out.jpg", 0)
# performing the segmentation on the input image with overlay
segment_image.segmentAsPascalvoc(img_file, output_image_name = "output_images/out_overlay.jpg", overlay = True)
out_overlay = plt.imread("output_images/out_overlay.jpg")
return out, out_overlay
def do_semantic(img_path):
absolute_path = os.path.abspath(
"model/deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
segment_image = semantic_segmentation()
# Load the xception model trained on pascal voc for segmenting objects.
segment_image.load_pascalvoc_model(absolute_path)
# segment_image.load_pascalvoc_model("pascal.h5")
# Perform segmentation on an image
segmap, segoverlay = segment_image.segmentAsPascalvoc(
img_path, output_image_name="image1_new.jpg", overlay=True)
return segoverlay
# output, segmap = segment_image.segmentAsPascalvoc("sample1.jpg")
# cv2.imwrite("img.jpg", segoverlay)
# print(segoverlay.shape)
Original file is located at
https://colab.research.google.com/drive/15lqIl-GIkW7G4QaYMcvYL4LgcPAChwTi
"""
##Definición de Librerias
import pixellib
from pixellib.semantic import semantic_segmentation
from pixellib.instance import instance_segmentation
##Semantic Segmentation
print("Semantic Image Segmentation Process...")
#Definition of semantic_segmentation in pixellib Library
segment_image = semantic_segmentation()
#Es necesario los pesos de deeplabv3_xception_tf_dim_ordering_tf_kernels.h5
segment_image.load_pascalvoc_model("deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
#Semantic Segmentation
segment_image.segmentAsPascalvoc("test7_4.jpg", output_image_name = "person_segmentationN.jpg")
#Semantic Segmentation with image Mask
segment_image.segmentAsPascalvoc("test7_4.jpg", output_image_name = "person_segmentationN_Mask.jpg", overlay = True)
print("Done")
##Instance Segmentation
print("Instance Image Segmentation Process...")
#https://github.com/ayoolaolafenwa/PixelLib/releases/download/1.3/deeplabv3_xception65_ade20k.h5
#https://github.com/ayoolaolafenwa/PixelLib/releases/download/1.1/deeplabv3_xception_tf_dim_ordering_tf_kernels.h5
# lets copy the files to c:/models
import pixellib
from pixellib.semantic import semantic_segmentation
import cv2
cam = cv2.VideoCapture(0)
# set the capturing resolotion to 1920 X 1080
cam.set(3, 1920)
cam.set(4, 1080)
semgment_video = semantic_segmentation()
#semgment_video.load_ade20k_model("c:/models/deeplabv3_xception65_ade20k.h5")
#seg, result = semgment_video.process_camera_ade20k(cam ,overlay=True, frames_per_second=15 , show_frames=True , frame_name="frame", output_video_name="c:/demo/cameraSemanticSegmentation.mp4" )
# remember to press q to quit the camera !!!
#lets use the second model
semgment_video.load_pascalvoc_model(
'c:/models/deeplabv3_xception_tf_dim_ordering_tf_kernels.h5')
seg, result = semgment_video.process_camera_pascalvoc(
cam,
overlay=True,
frames_per_second=15,
show_frames=True,
frame_name="frame",
def LoadSemanticSegmenter():
from pixellib.semantic import semantic_segmentation
global Segmenter_Semantic
Segmenter_Semantic = semantic_segmentation()
Segmenter_Semantic.load_pascalvoc_model(
ModelsDir + "deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
import cv2
#Pixellib is important library for Instance and Semantic segmentation
from pixellib.instance import instance_segmentation
from pixellib.semantic import semantic_segmentation
ins_seg = instance_segmentation() #For instance segementation
sem_seg = semantic_segmentation() #For semantic segementation
ins_seg.load_model('mask_rcnn_coco.h5')
#sem_seg.model('mask_rcnn_coco.h5')
cap = cv2.VideoCapture(0)
while True:
ret, frame = cap.read()
if ret:
result = ins_seg.segmentFrame(frame, show_bboxes=True)
#result = sem_seg.segmentFrameAsPascalvoc(frame)
#result is tuple containing dict in 0 index and image in 1 index
image = result[1]
cv2.imshow("Instance", image)
if cv2.waitKey(1) & 0xFF == ord("q"):
break
cap.release()
cv2.destroyAllWindows()
#SEMANTIC SEGMENTATION
import pixellib
from pixellib.semantic import semantic_segmentation
segment = semantic_segmentation()
segment.load_pascalvoc_model("pascal.h5")
out = segment.segmentAsPascalvoc("t1.jpg", output_image_name="image1.jpg")
"""#Apply segmentation overlay
import pixellib
from pixellib.semantic import semantic_segmentation
segment = semantic_segmentation()
segment.load_pascalvoc_model("pascal.h5")
out = segment.segmentAsPascalvoc("t1.jpg", output_image_name = "image2.jpg", overlay=True)
#INSTANCE SEGMENTATION
import pixellib
from pixellib.instance import instance_segmentation
instance_seg = instance_segmentation()
instance_seg.load_model("mask_rcnn_coco.h5")
out = instance_seg.segmentImage("img.jpg", output_image_name = "image3.jpg")
#Show segmented images with bounding boxes
import pixellib
from pixellib.instance import instance_segmentation
instance_seg = instance_segmentation()
instance_seg.load_model("mask_rcnn_coco.h5")
def __init__(self):
self.model = semantic_segmentation()
self.model.load_ade20k_model("deeplabv3_xception65_ade20k.h5")
import pixellib
from pixellib.semantic import semantic_segmentation
import cv2
#experimental script to play with better pixelators using various CV techniques. Sort of a mess right now
segment_frame = semantic_segmentation()
segment_frame.load_pascalvoc_model(
"deeplabv3_xception_tf_dim_ordering_tf_kernels.h5")
#segment_image.segmentAsPascalvoc("../images/person.jpg", output_image_name = "../images/image_new.jpg")
vid = cv2.VideoCapture(0)
while (True):
# Input image
#input = cv2.imread('../images/person.jpg')
ret, input = vid.read()
#input = cv2.imread('../images/person.jpg')
segmap, output = segment_frame.segmentFrameAsPascalvoc(input)
cv2.imshow('Input', input)
cv2.imshow('Output', output)
# the 'q' button is set as the
# quitting button you may use any
# desired button of your choice
if cv2.waitKey(1) & 0xFF == ord('q'):
break
import pixellib
#semantic segmentation
from pixellib.semantic import semantic_segmentation
segment_img = semantic_segmentation()
segment_img.load_ade20k_model("./model/deeplabv3_xception65_ade20k.h5")
segment_img.segmentAsAde20k("./img/1.jpg",
overlay=True,
output_image_name="sem_seg_res.jpg")
#instance segmentation
from pixellib.instance import instance_segmentation
segment_image = instance_segmentation()
segment_image.load_model("./model/mask_rcnn_coco.h5")
segment_image.segmentImage("./img/1.jpg",
show_bboxes=True,
output_image_name="ins_seg_res.jpg")
