Hari Mohan Pandey

Santosh Kumar Yadav, Achleshwar Luthra, Esha Pahwa et al.

Human activity recognition (HAR) using drone-mounted cameras has attracted considerable interest from the computer vision research community in recent years. A robust and efficient HAR system has a pivotal role in fields like video surveillance, crowd behavior analysis, sports analysis, and human-computer interaction. What makes it challenging are the complex poses, understanding different viewpoints, and the environmental scenarios where the action is taking place. To address such complexities, in this paper, we propose a novel Sparse Weighted Temporal Attention (SWTA) module to utilize sparsely sampled video frames for obtaining global weighted temporal attention. The proposed SWTA is comprised of two parts. First, temporal segment network that sparsely samples a given set of frames. Second, weighted temporal attention, which incorporates a fusion of attention maps derived from optical flow, with raw RGB images. This is followed by a basenet network, which comprises a convolutional neural network (CNN) module along with fully connected layers that provide us with activity recognition. The SWTA network can be used as a plug-in module to the existing deep CNN architectures, for optimizing them to learn temporal information by eliminating the need for a separate temporal stream. It has been evaluated on three publicly available benchmark datasets, namely Okutama, MOD20, and Drone-Action. The proposed model has received an accuracy of 72.76%, 92.56%, and 78.86% on the respective datasets thereby surpassing the previous state-of-the-art performances by a margin of 25.26%, 18.56%, and 2.94%, respectively.

Santosh Kumar Yadav, Apurv Shukla, Kamlesh Tiwari et al.

Pose recognition deals with designing algorithms to locate human body joints in a 2D/3D space and run inference on the estimated joint locations for predicting the poses. Yoga poses consist of some very complex postures. It imposes various challenges on the computer vision algorithms like occlusion, inter-class similarity, intra-class variability, viewpoint complexity, etc. This paper presents YPose, an efficient deep convolutional neural network (CNN) model to recognize yoga asanas from RGB images. The proposed model consists of four steps as follows: (a) first, the region of interest (ROI) is segmented using segmentation based approaches to extract the ROI from the original images; (b) second, these refined images are passed to a CNN architecture based on the backbone of EfficientNets for feature extraction; (c) third, dense refinement blocks, adapted from the architecture of densely connected networks are added to learn more diversified features; and (d) fourth, global average pooling and fully connected layers are applied for the classification of the multi-level hierarchy of the yoga poses. The proposed model has been tested on the Yoga-82 dataset. It is a publicly available benchmark dataset for yoga pose recognition. Experimental results show that the proposed model achieves the state-of-the-art on this dataset. The proposed model obtained an accuracy of 93.28%, which is an improvement over the earlier state-of-the-art (79.35%) with a margin of approximately 13.9%. The code will be made publicly available.

Santosh Kumar Yadav, Muhtashim Rafiqi, Egna Praneeth Gummana et al.

This paper presents a novel multimodal human activity recognition system. It uses a two-stream decision level fusion of vision and inertial sensors. In the first stream, raw RGB frames are passed to a part affinity field-based pose estimation network to detect the keypoints of the user. These keypoints are then pre-processed and inputted in a sliding window fashion to a specially designed convolutional neural network for the spatial feature extraction followed by regularized LSTMs to calculate the temporal features. The outputs of LSTM networks are then inputted to fully connected layers for classification. In the second stream, data obtained from inertial sensors are pre-processed and inputted to regularized LSTMs for the feature extraction followed by fully connected layers for the classification. At this stage, the SoftMax scores of two streams are then fused using the decision level fusion which gives the final prediction. Extensive experiments are conducted to evaluate the performance. Four multimodal standard benchmark datasets (UP-Fall detection, UTD-MHAD, Berkeley-MHAD, and C-MHAD) are used for experimentations. The accuracies obtained by the proposed system are 96.9 %, 97.6 %, 98.7 %, and 95.9 % respectively on the UP-Fall Detection, UTDMHAD, Berkeley-MHAD, and C-MHAD datasets. These results are far superior than the current state-of-the-art methods.

Santosh Kumar Yadav, Esha Pahwa, Achleshwar Luthra et al.

Drone-camera based human activity recognition (HAR) has received significant attention from the computer vision research community in the past few years. A robust and efficient HAR system has a pivotal role in fields like video surveillance, crowd behavior analysis, sports analysis, and human-computer interaction. What makes it challenging are the complex poses, understanding different viewpoints, and the environmental scenarios where the action is taking place. To address such complexities, in this paper, we propose a novel Sparse Weighted Temporal Fusion (SWTF) module to utilize sparsely sampled video frames for obtaining global weighted temporal fusion outcome. The proposed SWTF is divided into two components. First, a temporal segment network that sparsely samples a given set of frames. Second, weighted temporal fusion, that incorporates a fusion of feature maps derived from optical flow, with raw RGB images. This is followed by base-network, which comprises a convolutional neural network module along with fully connected layers that provide us with activity recognition. The SWTF network can be used as a plug-in module to the existing deep CNN architectures, for optimizing them to learn temporal information by eliminating the need for a separate temporal stream. It has been evaluated on three publicly available benchmark datasets, namely Okutama, MOD20, and Drone-Action. The proposed model has received an accuracy of 72.76%, 92.56%, and 78.86% on the respective datasets thereby surpassing the previous state-of-the-art performances by a significant margin.

Hari Mohan Pandey, Anshul Gupta, Subham Sarkar et al.

Text-to-SQL systems enable users to interact with structured databases using natural language, eliminating the need for specialized programming knowledge. In this work, we introduce GEMMA-SQL, a lightweight and efficient text-to-SQL model built upon the open-source Gemma 2B architecture. Unlike many large language models (LLMs), GEMMA-SQL is fine-tuned in a resource-efficient, iterative manner and can be deployed on low-cost hardware. Leveraging the SPIDER benchmark for training and evaluation, GEMMA-SQL combines multiple prompting strategies, including few-shot learning, to enhance SQL query generation accuracy. The instruction-tuned variant, GEMMA-SQL Instruct, achieves 66.8% Test-Suite accuracy and 63.3% Exact Set Match accuracy, outperforming several state-of-the-art baselines such as IRNet, RYANSQL, and CodeXDavinci. The proposed approach demonstrates that effective prompt design and targeted instruction tuning can significantly boost performance while maintaining high scalability and adaptability. These results position GEMMA-SQL as a practical, open-source alternative for robust and accessible text-to-SQL systems.

Jia Qian, Lars Kai Hansen, Xenofon Fafoutis et al.

Federated Learning is a framework that jointly trains a model \textit{with} complete knowledge on a remotely placed centralized server, but \textit{without} the requirement of accessing the data stored in distributed machines. Some work assumes that the data generated from edge devices are identically and independently sampled from a common population distribution. However, such ideal sampling may not be realistic in many contexts. Also, models based on intrinsic agency, such as active sampling schemes, may lead to highly biased sampling. So an imminent question is how robust Federated Learning is to biased sampling? In this work\footnote{\url{https://github.com/jiaqian/robustness_of_FL}}, we experimentally investigate two such scenarios. First, we study a centralized classifier aggregated from a collection of local classifiers trained with data having categorical heterogeneity. Second, we study a classifier aggregated from a collection of local classifiers trained by data through active sampling at the edge. We present evidence in both scenarios that Federated Learning is robust to data heterogeneity when local training iterations and communication frequency are appropriately chosen.

Hari Mohan Pandey

Large Language Models (LLMs) are transforming mental health care by enhancing accessibility, personalization, and efficiency in therapeutic interventions. These AI-driven tools empower mental health professionals with real-time support, improved data integration, and the ability to encourage care-seeking behaviors, particularly in underserved communities. By harnessing LLMs, practitioners can deliver more empathetic, tailored, and effective support, addressing longstanding gaps in mental health service provision. However, their implementation comes with significant challenges and ethical concerns. Performance limitations, data privacy risks, biased outputs, and the potential for generating misleading information underscore the critical need for stringent ethical guidelines and robust evaluation mechanisms. The sensitive nature of mental health data further necessitates meticulous safeguards to protect patient rights and ensure equitable access to AI-driven care. Proponents argue that LLMs have the potential to democratize mental health resources, while critics warn of risks such as misuse and the diminishment of human connection in therapy. Achieving a balance between innovation and ethical responsibility is imperative. This paper examines the transformative potential of LLMs in mental health care, highlights the associated technical and ethical complexities, and advocates for a collaborative, multidisciplinary approach to ensure these advancements align with the goal of providing compassionate, equitable, and effective mental health support.