○ Box Recognition using 3D cameras for Robot based pick and place
○ Successful Deployment to the customer warehouse
○ Developed Test Environment to evaluate system accuracy and version control
○ C++, Python, Opencv, PCL, tensorflow, mongodb, Django
○ Reflection Separation in Video : Extract Reflection and Background from video based on layer
motion and texture - Publication in ICCV2017
○ 2D to 3D conversion : Generate right view image from left image for 3D using fully
convolutional network
○ Single Image Facial Relighting : Build a CNN model to change lighting condition on facial
structure from single input image
○ MATLAB, Python, Tensorflow, matconvnet, caffe
○ Worked with online video streaming data
○ Develop Online Real-time Ads Detection Algorithms
○ Implemented Test System and Data Collection framework
○ FFMpeg, c++ tools, Linux system, AWS environment
○ Image Processing Algorithm Development
➢ 3D reconstruction of nano-scale structure from SEM images
➢ Detect Manufacturing Defect in Chip Layers
➢ Estimate Dimension variation of Design Patterns
➢ Mask Alignment Check
○
Accomplishments
➢ Idea Award for nano-scale 3D reconstruction project