I recently worked on a classical ill-posed problem of image colorization. We outperform most of the existing methods and achieved accuracy comparable to the state-of-the-art with a learning-based method. We are providing a parallel GAns training pipeline to handle the diverse data modes and cable of preserving the object-scene semantics to generate realistic plausible results. [Report]

Supervisor: Dr. Sanjay Singh 

Driving Assitant for Indian Roads: we worked on improving the driving assistant system for Indian road circumstances With a camera mounted on the vehicle dashboard. It entailed creating, optimizing, and adding Indian road events to the dataset for tracking multiple objects in a frame of the sequence using a deep learning approach. [Report ]

Supervisor: Prof. Brejesh Lall and Prof. Abhishek Gagneja

Drone-Based Crop Monitoring System: In which we proposed a crop health monitoring and visualization system based on drones. In this method, the agricultural field is covered by onboard cameras, which are used to evaluate and detect the unhealthy regions in the acquired images.  [Report ]

Supervisor: Prof. Abhishek Gagneja

AQI to Cloud Connectivity:  At Aerogram my work entailed building a python-based gateway from air monitoring devices to the cloud architecture and maintaining the real-time reading on the server side.

Advised by: Mr. Manoj Sahukar