We present the first unified framework for novel-view acoustic synthesis that entirely bypasses explicit 3D visual rendering and costly photogrammetry by directly grounding spatial audio generation in feed-forward visual geometry. Our framework synthesizes accurate and immersive spatial audio in 3D spaces without requiring viewpoint images, dense point maps, or ground-truth poses for input video. We propose the Geometry-Grounded Acoustic Decoder (GGAD) to dynamically attend to cross-modal features embedding local and global geometries in audio and visual modalities. Extensive experiments show that our framework outperforms prior work across various benchmarks in high-quality, viewpoint-accurate spatial audio synthesis, without requiring time-consuming explicit rendering of novel-view images or dense point maps.
@inproceedings{polra2026nvas,title={Visual Geometry Grounded Novel-View Acoustic Synthesis},author={Polra, Jay and Chauhan, Dhwanil and Huang, Wenjun and Toth, Kyle and Wang, Xianhui and Ni, Yang},booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},year={2026},url={https://dhwanil832.github.io/projects/nvas/},note={Purdue University Northwest · UC Irvine · San Diego State University}}
We present a multi-camera perception system for dynamic safety zone reconfiguration in active industrial environments. Static safety zone definitions fail in dynamic industrial floors where equipment and personnel positions shift continuously. Our system fuses four simultaneous camera perspectives using a rule-based spatial reasoning engine to redefine green zone boundaries in real time based on outermost detected blocker positions, enabling proactive hazard detection before incidents occur.
@inproceedings{toth2026safety,title={Development of Trialing Image Detection for a Melt Shop Safety Tool},author={Toth, Kyle and Polra, Jay and Chauhan, Dhwanil and Ni, Yang and Zhou, Chenn and Fisher, Conrad and Page, Garrett},booktitle={AISTech 2026 — Iron and Steel Technology Conference Proceedings},year={2026},url={https://dhwanil832.github.io/publications/},note={Purdue University Northwest · CIVS · SMSVC}}
The robustness of vision-language models under simultaneous visual and linguistic degradation — a common failure condition in real-world deployments — remains largely unstudied. We introduce a systematic benchmark evaluating 20 VLMs across controlled simultaneous multimodal corruption conditions, featuring a novel text corruption module and structured evaluation pipeline designed to characterize failure modes relevant to safety-critical and industrial deployment scenarios.
@article{chauhan2026vlmbenchmark,title={{VLM} Robustness Benchmark Under Simultaneous Multimodal Degradation},author={Chauhan, Dhwanil and others},journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},year={2026},url={https://dhwanil832.github.io/publications/},note={In Preparation — Targeting IEEE TPAMI / IJCV},}
We present the architectural evolution of an industrial safety AI system from a monolithic conversational pipeline to a modular multi-agent architecture, along with a benchmark for evaluating on-premise LLMs in agentic industrial deployment settings. The multi-agent design separates dialogue state management, context retention, action sequencing, and domain reasoning into specialized agents, enabling scalable and robust incident response in live manufacturing environments.
@inproceedings{chauhan2026aiaav2,title={{AI}-Assisted Accident Investigation {V2}: Multi-Agent Architecture Benchmark},author={Chauhan, Dhwanil and others},booktitle={Proceedings of the Annual Meeting of the Association for Computational Linguistics},year={2026},url={https://dhwanil832.github.io/publications/},note={In Preparation — Targeting ACL},}
@inproceedings{mithaiwala2026alzheimers,title={Early-Stage {Alzheimer's} Detection: Comparing {CNN} and Advanced Neural Architectures},author={Mithaiwala, Krish and Patel, Vidhiben Ka and Patel, Sachin and Chauhan, Dhwanil and Shah, Margi and Patel, Ankur},booktitle={ICT Systems and Sustainability},year={2026},pages={193--204},publisher={Springer Nature Switzerland},doi={10.1007/978-3-032-06674-9_20},url={https://doi.org/10.1007/978-3-032-06674-9_20},isbn={978-3-032-06674-9},note={CHARUSAT · DEPSTAR}}
@inproceedings{chauhan2026rfmal,title={{RF-MalDetect}: Harnessing Random Forest for Malware Identification in {PE} Files},author={Chauhan, Dhwanil and Patel, Sachin and Patel, Ankur and Patel, Rushi and Patel, Rishi and Bhatt, Rooshikesh and Shah, Margi and Seth, Manav},booktitle={SEET---Software Engineering for Emerging Technologies},year={2026},pages={3--22},publisher={Springer Nature Switzerland},doi={10.1007/978-3-032-08977-9_1},url={https://doi.org/10.1007/978-3-032-08977-9_1},isbn={978-3-032-08977-9},note={CHARUSAT · DEPSTAR}}
@inproceedings{patel2026watermarking,title={A Robust Parametric Watermarking Framework for Deep Neural Network Ownership Verification},author={Patel, Sachin and Patel, Ayush and Patel, Marmik and Soni, Rishi and Parmar, Selin and Patel, Premal and Chauhan, Dhwanil and Damodar, Dipika},booktitle={Soft Computing and Its Engineering Applications},year={2026},pages={455--466},publisher={Springer Nature Switzerland},doi={10.1007/978-3-032-22062-2_35},url={https://doi.org/10.1007/978-3-032-22062-2_35},isbn={978-3-032-22062-2},note={CHARUSAT · DEPSTAR — author name corrected from published typo "Dhawnil"}}
@inproceedings{savani2026lstm,title={A Comparative Study on Stock Market Prediction Using Long Short-Term Memory ({LSTM}) and Reinforcement Learning},author={Savani, Tilak A. and Patel, Sachin and Kathrotiya, Nishant and Chauhan, Dhwanil and Nayak, Amit and Patel, Premal},booktitle={World Congress on Smart Computing},year={2026},pages={303--315},publisher={Springer Nature Singapore},doi={10.1007/978-981-95-0183-0_25},url={https://doi.org/10.1007/978-981-95-0183-0_25},isbn={978-981-95-0183-0},note={CHARUSAT · DEPSTAR}}
Experimental evaluation demonstrates strong performance, achieving an F1 score of 0.92 for worker detection and 0.93 for personal protective equipment (PPE) classification. The integrated hazard-zone monitoring module further enables spatial safety enforcement, achieving a frame-level detection accuracy of 92% for restricted-area violations. The system supports real-time monitoring and processes video streams at approximately 35 ms per frame on a standard GPU. Preliminary deployment observations in an operational steel facility suggest that the system can assist safety personnel by improving visibility of PPE compliance and restricted-area monitoring. These results demonstrate the technical feasibility of the proposed modular framework as a scalable approach for automated industrial safety monitoring.
@article{Toth2026,author={Toth, Kyle and Singhal, Monika and Chauhan, Dhwanil and Polra, Jay and Zhou, Chenn and Page, Garrett and Fisher, Conrad},title={A Dual-Model Approach to Industrial Safety: Computer Vision for PPE Compliance and Hazard-Zone Monitoring in Steel Production},journal={Integrating Materials and Manufacturing Innovation},year={2026},month=may,day={12},issn={2193-9772},doi={10.1007/s40192-026-00461-6},note={Purdue University Northwest · CIVS · SMSVC},url={https://doi.org/10.1007/s40192-026-00461-6}}
Industrial safety incident reporting relies on manual form completion by on-shift supervisors immediately following high-stress events, producing inconsistent and incomplete records. We present an integrated conversational AI system for industrial safety incident management that guides supervisors through structured incident capture via natural dialogue, maintaining context across multi-turn exchanges to extract incident classification, risk assessment, root cause analysis, and corrective actions — replacing fragmented manual entry with consistent, structured safety records in live steel manufacturing environments.
@inproceedings{pu2025aiaai,title={Artificial Intelligence-Assisted Accident Investigation: Improving Safety Reporting with {LLMs}},author={Pu, Qingyun and Chauhan, Dhwanil and Moreland, John and Toth, Kyle and Zhou, Chenn and Archer, Erik and Berry, Nathan},booktitle={AISTech 2025 — Iron and Steel Technology Conference Proceedings},year={2025},url={https://dhwanil832.github.io/publications/},note={Purdue University Northwest · CIVS · SMSVC}}
@inproceedings{chauhan2025bags,title={Comparative Analysis of Deep Learning Models for Image Classification: A Study on Synthetic Images of Bags},author={Chauhan, Dhwanil and Patel, Sachin and Shah, Margi and Patel, Twisha and Sutariya, Sujal and Patel, Ankur and Patel, Diya and Patel, Isha},booktitle={ICT: Applications and Social Interfaces},year={2025},pages={323--336},publisher={Springer Nature Singapore},doi={10.1007/978-981-96-5754-4_28},url={https://doi.org/10.1007/978-981-96-5754-4_28},isbn={978-981-96-5754-4},note={CHARUSAT · DEPSTAR}}
@inproceedings{shah2024stress,title={Stress Detection Across Demographics: Leveraging Linear Regression Analysis},author={Shah, Kathit and Patel, Dhruvi and Chauhan, Dhwanil and Shah, Margi and Patel, Yash and Dubey, Nilesh and Patel, Sachin and Patel, Atul},booktitle={ICT for Intelligent Systems},year={2024},pages={503--516},publisher={Springer Nature Singapore},doi={10.1007/978-981-97-6675-8_42},url={https://doi.org/10.1007/978-981-97-6675-8_42},isbn={978-981-97-6675-8},note={CHARUSAT · DEPSTAR}}
@inproceedings{shah2024techtrends,title={Technological Trends and Their Impact on Society: A Comprehensive Analysis},author={Shah, Margi and Chauhan, Dhwanil and Patel, Sachin and Bhatt, Arpit and Patel, Chirag and Patel, Rushi and Patel, Ankur and Patel, Jay and Ramoliya, Dipak and Makwana, Hitesh and Nayak, Amit and Patel, Radhika and Jani, Ritika and Katira, Ashish and Patel, Rajesh and Sharma, Shital and Patel, Akash},booktitle={Proceedings of World Conference on Information Systems for Business Management},year={2024},pages={391--403},publisher={Springer Nature Singapore},doi={10.1007/978-981-99-8346-9_33},url={https://doi.org/10.1007/978-981-99-8346-9_33},isbn={978-981-99-8346-9},note={CHARUSAT · DEPSTAR}}
@inproceedings{chauhan2023iot,title={A Novel Intrusion Detection System based on Machine Learning for Internet of Things ({IoT}) Devices},author={Chauhan, Dhwanil and Shah, Margi and Joshi, Harshil},booktitle={2023 3rd International Conference on Smart Data Intelligence (ICSMDI)},year={2023},pages={427--434},doi={10.1109/ICSMDI57622.2023.00081},url={https://doi.org/10.1109/ICSMDI57622.2023.00081},note={CHARUSAT · DEPSTAR}}
@inproceedings{chauhan2023braintumor,title={Enhanced Brain Tumor Localization Techniques: A Paradigm Shift in Diagnosis},author={Chauhan, Dhwanil and Patel, Sachin and Shah, Margi and Bhatt, Arpit and Bhandari, Prakshal and Chauhan, Mehul},booktitle={2023 International Conference on Artificial Intelligence for Innovations in Healthcare Industries (ICAIIHI)},year={2023},volume={1},pages={1--7},doi={10.1109/ICAIIHI57871.2023.10489201},url={https://doi.org/10.1109/ICAIIHI57871.2023.10489201},note={CHARUSAT · DEPSTAR}}
@inproceedings{sahetai2023pulmonary,title={Harnessing Artificial Intelligence for Precise Pulmonary Disease Diagnosis},author={Sahetai, Muskan Omprakash and Patel, Sachin and Shah, Margi and Chauhan, Dhwanil and Patel, Radhika and Jani, Ritika},booktitle={2023 International Conference on Self Sustainable Artificial Intelligence Systems (ICSSAS)},year={2023},pages={151--157},doi={10.1109/ICSSAS57918.2023.10331667},url={https://doi.org/10.1109/ICSSAS57918.2023.10331667},note={CHARUSAT · DEPSTAR}}
@inproceedings{bhatt2023ethereum,title={Ethereum Blockchain Wallets: A {Kotlin}-based Implementation Perspective},author={Bhatt, Arpit and Chauhan, Dhwanil and Shah, Margi and Patel, Sachin and Sudani, Jay and Vachhani, Krish},booktitle={2023 7th International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)},year={2023},pages={128--134},doi={10.1109/I-SMAC58438.2023.10290281},url={https://doi.org/10.1109/I-SMAC58438.2023.10290281},note={CHARUSAT · DEPSTAR}}