Amherst, New York United States
Principal Software Engineer
Technology / Software / Internet
Goals: Practical large scale content analysis
Professional Experience:
Video to Text Translation without using Expensive Frame-by-frame Manual Annotation:
• Innovative formulation of unsupervised probabilistic latent variable models that allow discovery of topics from documents containing real life videos and very short summaries
• Extensive experimentations with topic models that translate related words to video frames and also translate a new video to relevant keywords and then to natural language phrases and sentences using language models and parse trees
• Provided solutions for scenarios where manual frame-by-frame annotation of real life longer videos are difficult and no