try: import cloudinary except: import os os.system('pip install cloudinary') try: import detectron2 except: import os os.system('pip install git+https://github.com/facebookresearch/detectron2.git') import cv2 import json from matplotlib.pyplot import axis import gradio as gr import requests import numpy as np from torch import nn import requests from numpy.lib.type_check import imag import random import time import csv import torch from detectron2 import model_zoo from detectron2.engine import DefaultPredictor from detectron2.config import get_cfg from detectron2.utils.visualizer import Visualizer from detectron2.data import MetadataCatalog from detectron2.data.datasets import register_coco_instances # import some common detectron2 utilities from detectron2.engine import DefaultPredictor from detectron2.config import get_cfg from detectron2.utils.visualizer import Visualizer from detectron2.data import MetadataCatalog, DatasetCatalog from detectron2.utils.visualizer import ColorMode import cloudinary.uploader import io import os API_KEY =os.environ["API_KEY"] Cloud_Name =os.environ["Cloud_Name"] API_Secret =os.environ["API_Secret"] cfg = get_cfg() # try: # register_coco_instances("Fiber", {}, "./labels-fiver.json", "./Fiber") # Fiber_metadata = MetadataCatalog.get("Fiber") # dataset_dicts = DatasetCatalog.get("Fiber") # except: # print("there is an issue") register_coco_instances("Fiber", {}, "./labels-fiver.json", "./Fiber") Fiber_metadata = MetadataCatalog.get("Fiber") dataset_dicts = DatasetCatalog.get("Fiber") model_path = "./model_final.pth" # cfg = get_cfg() # cfg.merge_from_file("./configs/detectron2/faster_rcnn_R_50_FPN_3x.yaml") cfg.MODEL.ROI_HEADS.NUM_CLASSES = 2 # cfg.MODEL.WEIGHTS = model_path my_metadata = MetadataCatalog.get("dbmdz_coco_all") # Fiber_metadata.thing_classes = ["Fiber", "Fiber","Fiber"] my_metadata.thing_classes = ["Fiber", "Fiber"] cfg.merge_from_file("./configs/detectron2/mask_rcnn_R_50_FPN_3x.yaml") cfg.MODEL.WEIGHTS = model_path #os.path.join(cfg.OUTPUT_DIR, "model_final.pth") cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.05 # set the testing threshold for this model cfg.DATASETS.TEST = ("fiber", ) # predictor = DefaultPredictor(cfg) if not torch.cuda.is_available(): cfg.MODEL.DEVICE = "cpu" def inference(image_url, image, min_score): if image_url: r = requests.get(image_url) if r: im = np.frombuffer(r.content, dtype="uint8") im = cv2.imdecode(im, cv2.IMREAD_COLOR) else: # Model expect BGR! im = image[:,:,::-1] cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.005 predictor = DefaultPredictor(cfg) outputs = predictor(im) # v = Visualizer(im, my_metadata, scale=1) # out = v.draw_instance_predictions(outputs["instances"].to("cpu")) # for d in random.sample(dataset_dicts, 3): # img = cv2.imread("https://meet.google.com/ice-wndh-joi.jpg") # predictor(img) # visualizer = Visualizer(img[:, :, ::-1], metadata=Fiber_metadata, scale=1) # vis = visualizer.draw_dataset_dict(d) # cv2_imshow(vis.get_image()[:, :, ::-1]) # !zip -r ./fiber.zip "/content/Fiber" # import json # from google.colab import files # uploaded = files.upload() # im = cv2.imread([key for key in uploaded.keys() # ][0]) # im = cv2.imread(d["file_name"]) # outputs = predictor(im) v = Visualizer(im[:, :, ::-1], metadata=my_metadata , #Fiber_metadata, scale=1, instance_mode=ColorMode.IMAGE_BW # remove the colors of unsegmented pixels ) v = v.draw_instance_predictions(outputs["instances"].to("cpu")) # cv2_imshow(v.get_image()[:, :, ::-1]) # print(outputs["instances"]) masks = np.asarray(outputs["instances"].pred_masks.to("cpu")) # bbox = np.asarray(outputs["instances"].pred_boxes.to("cpu")) # Pick an item to mask # img=v.get_image() # Define a dictionary to store the measurements and their positions measurements = {} for ind,item_mask in enumerate(masks): segmentation = np.where(item_mask == True) if segmentation[1].any() and segmentation[0].any(): # box=bbox[ind] # Get the true bounding box of the mask (not the same as the bbox prediction) x_min = int(np.min(segmentation[1])) x_max = int(np.max(segmentation[1])) y_min = int(np.min(segmentation[0])) y_max = int(np.max(segmentation[0])) measurement = int(0.5+len(segmentation[0])/600) measurements[ind] = {'measurement': measurement, 'x_min': x_min, 'x_max': x_max, 'y_min': y_min, 'y_max': y_max} # cv2.putText(img=img, text=str(int(0.5+len( segmentation[0])/600)), org=(x_min+20,y_min-10), fontFace=cv2.FONT_HERSHEY_TRIPLEX, fontScale=0.8, color=(0, 255, 0),thickness=2) # Loop over the masks # return outputs for ind, item_mask in enumerate(masks): segmentation = np.where(item_mask == True) measurement = int(0.5+len(segmentation[0])/600) measurements[ind] = {'measurement': measurement, 'x_min': x_min, 'x_max': x_max, 'y_min': y_min, 'y_max': y_max} cloudinary.config( cloud_name = Cloud_Name, api_key = API_KEY, api_secret = API_Secret, secure = True ) # Replace with the name of your CSV file filename = "your_file.csv" # Write the measurements to a CSV file filename=str(time.time())+'dmeasurements.csv' with open(filename, mode='w') as file: writer = csv.writer(file) writer.writerow(['ID', 'Measurement', 'X_Min', 'X_Max', 'Y_Min', 'Y_Max']) for id, data in measurements.items(): writer.writerow([id, data['measurement'], data['x_min'], data['x_max'], data['y_min'], data['y_max']]) # Convert the CSV content to a bytes object csv_bytes = io.StringIO( open(filename,"r").read()).read().encode("utf-8") # Upload the file to Cloudinary upload_result = cloudinary.uploader.upload( csv_bytes, resource_type = "raw", folder = "csv_files", public_id =filename, overwrite = False ) # return file return upload_result["url"], v.get_image() title = " fi ber detec tion Model " description = "" article = '' gr.Interface( inference, [gr.inputs.Textbox(label="Image URL", placeholder="https://api.digitale-sammlungen.de/iiif/image/v2/bsb10483966_00008/full/500,/0/default.jpg"), gr.inputs.Image(type="numpy", label="Input Image"), gr.Slider(minimum=0.0, maximum=1.0, value=0.01, label="Minimum score"), ], title=title, description=description, article=article, examples=[], outputs=[ "text","image"], ).launch()