MEDIA ANALYTICS
PROJECTS
PEOPLE
PUBLICATIONS
PATENTS
Media Analytics
Sparsh Garg is a Senior Associate Researcher in the Media Analytics Department at NEC Laboratories America. He received his MS in Computer Science and Engineering from Santa Clara University and his BTec in Electronics and Communications Engineering from Manipal Institute of Technology.
At NEC, his research focuses on leveraging LLMs and VLMs to design autonomous AI agents capable of orchestrating and automating the full lifecycle of AI models, from data processing to deployment. He designs efficient and scalable learning algorithms for visual AI, with applications in object detection, person re-identification, video summarization, and perception models for complex, dynamic environments.
His work addresses critical challenges such as adapting to new domains, learning from diverse and often conflicting datasets, and handling rare or previously unseen objects—capabilities essential for ensuring AI reliability in real-world deployments. He develops methods that integrate semantic segmentation, multi-modal learning, and architecture optimization to create models that are not only high-performing but also computationally efficient and interpretable. This balance allows NEC’s visual AI platforms to deliver accurate, explainable results in time-sensitive and safety-critical contexts, from autonomous systems to large-scale video analytics. By combining deep technical innovation with practical deployment considerations, his research strengthens NEC’s ability to deliver AI solutions that operate robustly across varied conditions and evolving operational environments.
Overview: Our AI DevOps pipeline builds a high-fidelity digital twin of sensor data which allows for self-improvement of deployed models. We leverage our foundational vision-language models to automatically determine issues in currently deployed AI, pseudo-label or simulate training data, develop models with continual learning and use an LLM-based verification over diverse scenarios.
Overview: Our techniques from unsupervised and semi-supervised learning, such as domain adaptation and domain generalization, allow robust and responsible AI solutions across multiple applications such as image classification, face recognition, facial anti-spoofing, object detection and semantic segmentation.
Video retrieval at scale is central to data curation and safety validation in autonomous driving, where users want to find not only scenes but also dynamic events such as cut-ins and hard braking. Existing vision-language and keyword-based retrieval methods often miss these events because the relevant
Driving vision-language models (VLMs) must accurately understand scenes across diverse conditions defined by Operational Design Domains (ODDs), yet verificationremains sparse: many slices are missing, making empirical failure rates unreliable. We propose SLICESCORER, a deterministic scoring rule for
Grounding large language models (LLMs) in domain-specific tasks like post-hoc dash-cam driving video analysis is challenging due to their general-purpose training and lack of structured inductive biases. As vision is often the sole modality available for such analysis (i.e. no LiDAR, GPS, etc.), existing
This paper proposes AutoScape, a long-horizon driving scene generation framework. At its core is a novel RGB-D diffusion model that iteratively generates sparse, geometrically consistent keyframes, serving as reliable anchors for the scenes appearance and geometry. To maintain long-range geometric consistency,
Obtaining high-quality fine-grained annotations for traffic signs is critical for accurate and safe decision-making in autonomous driving. Widely used datasets, such as Mapillary, often provide only coarse-grained labels without distinguishing semantically important types such as stop signs or speed
Our Sparsh Garg, a Senior Associate Researcher in the Media Analytics Department, will present “Mapillary Vistas Validation for Fine-Grained Traffic Signs: A Benchmark Revealing Vision-Language Model Limitations” at the Data Computer Vision (DataCV) 2025 workshop as part of ICCV 2025 in Honolulu,
Grounding large language models (LLMs) in domain-specific tasks like post-hoc dash-cam driving video analysis is challenging due to their general-purpose training and lack of structured inductive biases. As vision is often the sole modality available for such analysis (i.e., no LiDAR, GPS, etc.), existing
Autonomous vehicle (AV) systems rely on robust perception models as a cornerstone of safety assurance. However, objects encountered on the road exhibit a long-tailed distribution, with rare or unseen categories posing challenges to a deployed perception model. This necessitates an expensive process of
While it is desirable to train segmentation models on an aggregation of multiple datasets, a major challenge is that the label space of each dataset may be in conflict with one another. To tackle this challenge, we propose UniSeg, an effective and model-agnostic approach to automatically train segmentation
Test-time adaptation approaches have recently emerged as a practical solution for handling domain shift without access to the source domain data. In this paper, we propose and explore a new multi-modal extension of test-time adaptation for 3D semantic segmentation. We find that, directly applying existing
