Image Annotation / Labelling

We are looking for freelancers to accurately draw rectangles around all the food-products in the images and label it. We also have a pre-labelled dataset wherein certain images need to be cleared off and relabeled. The possible labels are as follows: 7_Up, Amul_Buttermilk, Bisleri_B, Blue_Lays, Coca_Cola, Maggie_Cuppa, Nescafe_Cl, Pepsi_Can, Pink_Kettle, Rajma_Rice, Sunfeast_B, Tata_Biryani, Too_Yumm, Cadbury_Silk, Bourbon. The rectangles around the products must be as close to the contour of the product as possible. In case of certain products, we will advise what part (top/side/bottom view) of the product needs to be labelled. New Images to label: 1000 Cleaning and editing labels: 3700 images. We will verify the quality before sending a new task. Before bidding check the annotation tool: 1) Download labelImg - [login to view URL] 2) Download/Clon the git repo for labelImg API 3) Go into the 'labelimg/data' directory and edit the '[login to view URL]' file with our custom classes 4) Run [login to view URL] from command line 5) On labelImg window enable 'Autosave' from 'View Menu' 6) Select 'Pascal VOC' format For more help: [login to view URL] 7) Once you click open dir, all images present in that directory will be loaded into labelImg 8) Select a directory to save the encoded .xml files. 9) Draw rectangles around every object in the image and selct its class and click next image. The .xml file will be autosaved. 10) Please check this for HotKeys and additional info --> [login to view URL] Few images from the dataset are enclosed in the attached file.