|
Ongoing Projects
| |
|
Swarath: autonomous vehicle of IIITD
(Details coming soon)
Querying in Camera Networks for Efficient Target Tracking
Team: Anil Sharma, Saket Anand and Sanjit K. Kaul.
(Details coming soon)
Completed Projects
| |
|
Reinforcement Learning Based Querying in Camera Networks for Efficient Target Tracking
Anil Sharma, Saket Anand, Sanjit Kaul
The volume of data generated by camera networks is humongous, and scalable processing of this data is imperative for deploying practical tracking of target.
In this paper, we address the largely overlooked problem of scheduling cameras for processing by selecting one where the target is most likely to appear next.
The inter-camera handover can then be performed on the selected cameras via re-identification or another target association technique.
We model this scheduling problem using reinforcement learning and learn the camera selection policy using Q-learning.
[PDF] [Github] [Video]
|
| |
|
|
Smartphone Audio based Distress Detection
Anil Sharma, Sarthak Ahuja, Mayank Gautam, Sanjit Kaul
In this project, we propose a novel two-stage supervised learning based method for monitoring signs of distress by leveraging the microphone of a smartphone. It is motivated by the possibility of 24x7 monitoring of a human about his daily routing using a smartphone. We investigate the signs of distress (screaming and crying sounds) in urban environment and it was evaluated over hundred hours of audio data recorded by 10 volunteers as they went about their daily routine. Interested users can refer to our paper "Two-Stage Supervised Learning-Based Method to
Detect Screams and Cries in Urban Environments" for more details.
[PDF] [Code] [more details]
|
| |
|
 |
BlindAssist: Nearby Building Information Advisory
Alvika Gautam, Anil Sharma, Lokender Tiwari, Chetan Arora
BlindAssist provides assistance to a blind person in finding a multi-storey building using a video captured from his smartphone. The video captured from the mobile phone is processed using sophisticated computer vision algorithm to detect number of floors in a building, if available in the video frames. The system then reports the distance of the building from the user. Interested users can refer to the report for more details.
[pdf]
|
| |
|
 |
Vehicle theft warning and tracking system
Anil Sharma, Samit Anwer, Vinayak Naik
Worrying about vehicle theft? No problem, this app may allow you to track your vehicle. Unlike other conventional systems, the owner (you) doesn't have to reload a webpage to keep track of the vehicle's location rather the system deployed in the vehicle will notify when its owner is not in proximity. The proximity is checked using GPS location of the car and its owner. If the vehicle is found to be outside the proxity region then the system inside vehicle will send a cellular message to the owner asking to confirm a theft, if the owner confirms theft then vehicle will continuously update its location to a server which can also be portrayed in the owner's android app. Currently we have smartphone app for both owner and vehicle.
|
| |
|
 |
keepQuiet: Noise level warning using temporal thresholding for shared space
Sarthak Ahuja, Anil Sharma
Frustrated from annoying labmates? If the answer is yes then this project may help you notify them and to automatically ask to keep quiet. Problem solved :), isn't?. Well this is a smartphone app that detects the noise level inside a room or say inside your lab. The novelity is not the detection rather selection of threshold for raising a notification, the threshold is a function of both sound intensity and time duration. This could benefit people who want to respect all in a shared space.
|
| |
|
Publications
- Anil Sharma, Saket Anand, Sanjit Kaul "Intelligent Camera Selection Decisions for Target Tracking in a Camera Network" accepted in WACV, 2022. [PDF]
- Anil Sharma, Saket Anand, Sanjit Kaul "Intelligent Querying for Target Tracking in Camera Networks using Deep Q-Learning with n-Step Bootstrapping" accepted in Image and Vision Computing, 2020. [PDF]
- Anil Sharma, Mayank K Pal, Saket Anand, Sanjit Kaul "Stratified Sampling Based Experience Replay for Efficient Camera Selection Decisions" accepted in BigMM 2020. [PDF]
- Anil Sharma, Saket Anand, Sanjit Kaul "Reinforcement Learning Based Querying in Camera Networks for Efficient Target Tracking" accepted for publication in ICAPS 2019. [PDF][Video]
- Mayank Pal, Rupali Bhati, Anil Sharma, Sanjit Kaul, Saket Anand, PB Sujit "A Reinforcement Learning Approach to Jointly Adapt Vehicular Communications and Planning for Optimized Driving" accepted for publication in ITSC 2018.
(pdf).
- Anil Sharma, Prabhat Kumar, Saket Anand, Sanjit Kaul, "A Reinforcement Learning Approach to Target Tracking in a Camera Network" accepted for presentation in Cognitive Vision: Intregrated Vision and AI for Embodied Perception and Interaction, Workshop with ICML/IJCAI-ECAI 2018.
(pdf).
- Anil Sharma, AB Buduru, "Foresee: Attentive Future Projections of Chaotic Road Environments with Online Training" in Proceedings of AAMAS 2018 (accepted as extended abstract), Stockholm, Sweden. (pdf).
- Anil Sharma, Sanjit Kaul, "Two-Stage Supervised Learning-Based Method to
Detect Screams and Cries in Urban Environments" published in IEEE/ACM transactions on Audio, Speech and Language Processing,
2016 (IEEE).
Awards & Achievements
- Received IJCAI conference travel grant to attend IJCAI-2019 in Macau.
- Received Google travel grant to attend ICAPS-2019 in Berkeley, CA, USA.
- Qualified NET (National Eligibility Test) with 99.77 percentile for JRF conducted by UGC
, Govt. of India.
- Received ICML student travel grant to attend ICML-2018 in Stockholm, Sweden.
- Received Microsoft Travel Grant to attend AAMAS-2018 in Stockholm, Sweden.
- Top Writer Tag on Quora for my answers on Reinforcement learning and Deep learning.
- First prize in Elevator Pitch at Research Showcase 2015, IIIT-Delhi .
Presentations, Lectures & Talks
- [PPT] Tutorial on basics of Probability and Statistics, IIIT-Delhi .
- [PPT] Unsupervised Classification: lecture in Machine Learning course in Summer 2014 , IIIT-Delhi .
- Instructor in summer refresher module on Introduction to C, Summer 2014. , IIIT-Delhi .
Databases, Codes & Tutorials
Other Useful Information:
Development Tools :
- Speech Processing: Praat, NSL
- Computer Vision: OpenCV (Android & C++), VLFeat
- Matlab (expert), Android
Online presentation:
|
