Partha Pratim Mohanta
Papers
1
Total Citations
3
H-Index
1
About
Partha Pratim Mohanta is a researcher specializing in computer vision and video understanding, with a particular focus on deep learning architectures for spatiotemporal data. His work addresses critical challenges in processing untrimmed, real-world videos captured from diverse sources such as social media, robotics, and surveillance systems. Mohanta’s key contribution lies in developing an unsupervised action localization crop method for video retargeting, designed to optimize input for 3D Convolutional Neural Networks (ConvNets). This approach solves the common problem of aspect ratio variability in videos, where traditional random- or center-cropping often discards the main subject. By intelligently localizing and cropping actions without requiring labeled data, his method enhances the efficiency and accuracy of 3D ConvNets, making them more robust for practical applications. Though his most-cited paper has garnered 3 citations to date, the work represents an important step toward automated video preprocessing. Mohanta’s research is particularly relevant for advancing video analysis in autonomous systems and content understanding, where handling unconstrained footage is essential.
Research Focus
Key Achievements
Top Papers
- 1