Dec 2016 – Jan 2019 · 2 yrs 1 mo · Bengaluru, Karnataka, India

• Tools/Technologies: Deep Learning, Computer Vision, iOS, Android, Python
• Project: Scanner SDK for Expense Reader
• User scans documents/receipts using iOS/Android mobile app. The system, in turn, finds document boundaries using a Deep Learning based solution. As a fall-back, interface is provided to edit document boundaries, if system doesn't crop the boundaries properly. Also, as we don't have control over whether user scans single document or multiple documents and/or tangible or intangible documents, deep learning based solutions are developed and integrated to guide the user.

Aug 2014 – Nov 2016 · 2 yrs 3 mos · Bangalore

• Summary: Worked on Computer Vision applications such as depth estimation systems (Computer Vision), Automated Optical Inspection (AOI) systems, Automatic Final Verification & Inspection (AFVI) systems
• Tools/Technologies: Image Processing, Computer Vision, C/C++, Android Porting, CUDA, SSE2, OpenCL
• Project1: Multi-shot (Action Sequence Imaging) App
• It aims at devising an algorithm that can convert different actions of an object in a video into single image, which can provide better resolution lower noise, much lesser storage space requirements than video.
• Project2: Low Light Image Enhancement and Outfocus using Dual Camera
• It aims at devising a method to find depth of objects, which needs to be highly accurate and should find depth of objects for high resolution images, in the surroundings for mobile applications. Disparity is used to enhance low-quality right image. Depth can be used to improve accuracy of disparity and thus helps in enhacing outfocus accuracy.
• Project3: BGA Color AFVI
• It aims at devising an algorithm for thin pattern creation and optimization of colour image warping
• Project4: SEMIPL
• It aims at devising a library for image processing. My contribution was to optimize morphological operations on CPU.
• Project5: GPCAMM
• It aims at developing ODB file based AOI systems to support PCB manufacturing. My role was to optimize line segmentation using OpenCL

Jun 2010 – Jul 2014 · 4 yrs 1 mo · Pune

• Summary: Worked on Driver Assistance Systems, Image Compression, Video Surveillance Systems
• Tools/Technologies: Image Processing, Computer Vision, MATLAB, C/C++, OpenCV
• Project1: Frugal Replacement for LiDAR
• It aims at devising a cost-effective method to find depth of objects in the surroundings for an autonomous vehicle
• Project2: Multi-camera Surveillance System
• This system is used to monitor a large area in order to track moving objects across multiple cameras, and thus making the system able to compose the unusual incidents (such as leaving bags at public places) automatically.
• Project3: 2D-3D conversion
• This system aims at estimating depth profile of faces and thus face features are extracted and saved in the database. This can be used for authentication purpose in security systems.
• Project4: Driver Status Monitoring
• Project5: Lossless Image Compression based on Data Folding

Jan 2009 – Jan 2009 · 0 mo

Aug 2006 – May 2010 · 3 yrs 9 mos · Greater Jabalpur Area

• Computer Science & Engineering