50 %
50 %
Information about Deepashree_Resume

Published on January 4, 2017

Author: DeepashreeG


1. 1 DEEPASHREE GURUMURTHY 6689 El Colegio Road, Apt 137 Phone: (805)-886-0876 Goleta, CA – 93117,USA email: Visa status: F1 Objective : Seeking a Full Time position to work on problems in field of Image Processing and Computer Vision Technical skills : MATLAB, Python, C/C++, OpenCV, Deep Learning, Object Recognition, NI LabVIEW Education : - September 2015 – Present: University of California, Santa Barbara, CA MS in Electrical Engineering with emphasis in Image Processing and Computer Vision, Current GPA: 3.70 Courses – Computer Imaging, Advanced topics in Computer Vision, Imaging Systems, Digital Image Processing, Pattern Recognition, Matrix analysis and computations, Stochastic Processes in engineering, Robotics- Haptics, Digital Speech Processing, Advanced Digital Signal Processing - September ’11 – June ’15: Visvesvaraya Tech. University, India. B.Eng in Electronics and Communication Engineering., CGPA: 4 (84% eqvt.) Experience : - October 2016 – Present: Visual Search and Object Recognition Intern at MAV Farm. Assessing the feasibility of object recognition and visual search within the scope of the company. Projects : 1. December 2016: HoloSwap: Object Removal and Replacement HoloSwap is an application on Microsoft Hololens that allows users to select an object of interest and removes or replaces it with another object in real time. 2. July 2016 (ongoing): Photo Summarization of Cities Involves extraction of mid-level representative patches from geotagged input images using deep learning and summarization of a region by using one discriminative exemplar image. 3. June 2016: Object recognition using bag of visual words A histogram of occurrences of visual words or bag of features is generated using k-means clustering algorithm on feature descriptors from input images of each category and is used to train images . A new image is then categorized based on the new classifier. 4. June 2016: Image Forensics A forensics test to determine whether the test image was tampered. Used various algorithms such as Copy-move algorithm, Error Level Analysis and explored deep learning methods. 5. May 2016: Structure from Motion using Two Views Sparse, feature based 3D reconstruction algorithm of two images of the same object clicked at different camera angles. Implemented on MATLAB. 6. May 2016: JPEG Compression Implementation of JPEG standard for lossy data compression and lossless data compression using Huffman encoding on MATLAB 7. May 2016: Multi-frequency Ground Penetration Radar Imaging Image reconstruction of rebars present below ground by plane to plane backpropagation and range profiles. Implemented on MATLAB.

2. 2 8. May 2016: Panorama Stitching Stitching of multiple related images to create a panorama. Implemented on MATLAB. 9. April 2016: Lip Tracking using Snake Model Used active contour model for lip tracking in series of frames from a video sequence. With an appropriate template, almost accurate movements to nearby edges and lip contour marking were achieved. Implemented on MATLAB. 10. April 2016: Image Registration Alignment of two or more images of same scene by applying geometric transformations, SIFT and RANSAC algorithms. Implemented on MATLAB. 11. February – March 2016: LPC Analysis and Synthesis of speech signals To obtain the reconstruction of speech using parameters extracted from a speech signal by developing algorithms to determine voiced and unvoiced segments of speech, pitch period, gain and design of filter using LPC coefficients. Implemented on MATLAB. 12. February – March 2016: Haptic Navigator: A wearable haptic feedback device Involved developing a wearable device (sensory substitution device) with vibrotactile haptic feedback by mapping space to time-delay to aid the visually impaired. 13. November 2014 – May 2015: Implementation of Adaptive Rate Compressive Sensing for background subtraction in a video sequence (TCS innovation labs, Bangalore, India) Achieved successful reconstruction of the foreground using fewer measurements than the conventional method of Nyquist sampling by varying the number of measurements taken in every frame using side information. 14. September - October 2013: Motion detection by image processing on MATLAB A simple motion detection algorithm implemented in Real Time. Achievements : • Ranked first in class at Brigade School in the annual examination of Central Board of Secondary Education, 2009 • Secured Awards for instrumental music and represented college in a talent tour for the same Interests : Photography, Playing piano, Music, Camping, Touring

Add a comment