Jewellery segmentation

Position - Project Assistant at Multimodal Perception Laboratory, International Institute of Information Technology
Technical skills - C++, OpenCV

In digital image processing and computer vision, image segmentation is a process of partitioning a digital image into multiple segments. It helps to simplify the image as well as the partitioned image is eventually assigned labels which with further processing play an important role in object detection within an image.
  I worked as a Project Assistant at Multimodal Perception Laboratory, International Institute of Information Technology, Bangalore, India. My work here was to analyze different image segmentation algorithms and develop an algorithm that would segment the background in shared jewelry images. The output of our algorithm was to replace the original background with transparent background. Our vendors intended to replace the background, edit the image and use it on selling websites.
  Figures 1 and 2 displays our algorithm’s performance on original images shared by our vendors.

jew1
Figure 1: Segmentation algorithm (best case)
jew2
Figure 2: Segmentation algorithm (worst case)
Dimple Bhuta
Dimple Bhuta
Principal Engineer – State Estimation and Control

My research interests include robotics, computer vision and bio-inspired design.