No ways rapidly rising crime rates due to digital fraud and which is limited access to real-time environmental updates, group of users struggle to stay informed and get quick assistance through various emergencies situations. This paper mainly focus on develops an intelligent, secure interactive chatbot to support strengthen for public safety for safely livelihood and digital crime , fraud awareness. This awareness acting as a digital link up between users and law enforcement securely, the chatbot identifies to provides local crime statistics with auto updates, practical safety guidelines, and access to essential support customer support services. It majorly queries a controlled crime dataset to assist user requests and helps users understand nearby current crime patterns and trends. Using NLP and ML methods, the proposed system identifies user intent and key challenging entities to generate exact accurate, conversational replies based on situational demands. The designed chatbot is wisely neural network model for query deep classification and NLTK for text cleaning, image and preprocessing. It also supports additionally online complaint registration, confirmations nearby police stations through an interactive map connecting platforms, and includes an SOS alert feature for urgent situations using Traffic Management System to it improves user accessibility, awareness, and communication for safer communities.

Keywords : Public Safety Chatbot, Crime Awareness System , NLP, Neural Network–Based Query Classification

Authors : D. Aruna Kumari

Title : CIVIC EYE : AI - Powered Public Safety Assistant

Volume/Issue : 2026;3(1 (January - March))

Page No : 86 - 90

The rapid development of 6G networks requires a communication system that can interpret and convey the hidden meaning of multimodal data rather than raw signals. Semantic communication is needed because it helps systems recognize and transmit the intended meaning of multimodal data. The current paper proposes a new model, the Advanced Multi-Modal Semantic Communication (AIMMSC) framework, for image-based semantic communication only. The suggested system relies on a hybrid ResNet-ViT architecture to detect and encode high-level visual semantics into a compact latent space, enabling resilient and efficient image communication over a low-bandwidth, noisy channel via a contrastive learning-powered semantic space with attention and transformers. AIMMSC's performance in eight key areas, including semantic accuracy, latency, robustness to noise, cross-modal alignment, adaptability, energy efficiency, and scalability, was compared with five new semantic communication models: DeepSC, Semantic-SC, CLIP-Comm, Auto-SC, and Transformer-SC. Experimental results show that IMMSC outperforms other approaches, improving semantic correctness by 19%, robustness by 22%, and cross-modal alignment by an average of 27%. The integration of several modalities into a single latent space and the use of deep semantic encoding are the sources of these developments. Such an approach lays the foundation for intelligent, context-aware communication. In next-generation networks, it dramatically improves the effectiveness and applicability of data interchange. For 6G applications like remote healthcare,extended reality (XR), and driverless cars, this foundation is essential.

Keywords : 6G communication, Advanced Multi-Modal, Semantic Communication, ResNet-ViT, Deep Learning.

Authors : T Venkata Krishnamoorthy , M Dharani , Ch Vijayalakshmi , N Subba Rayudu

Title : Advanced Multi-Modal Semantic Communication Using Hybrid ResNet-ViT Architecture for 6G Applications

Volume/Issue : 2026;3(1 (January - March))

Page No : 74 - 85

Kidney diseases such as cysts, stones, and tumors are a major global health burden, often remaining undiagnosed until advanced stages due to limited access to specialized care. While deep learning models like YOLOv8 have shown promise in automating detection from medical images, their performance is often hindered by class imbalance and poor feature discrimination in complex clinical data. This paper introduces an enhanced deep learning framework for the accurate multi-class classification of kidney abnormalities from computed tomography (CT) scans. Our novel approach integrates a Convolutional Block Attention Module (CBAM) into the YOLOv8 architecture, enabling the model to focus diagnostically relevant regions within the kidney. To address significant class imbalance, we adopt a Focal Loss function, which prioritizes difficult and underrepresented cases during training. We further enhance model robustness through an advanced data augmentation pipeline incorporating mixup and cutmix strategies. The proposed system is trained and evaluated on a publicly available dataset of 12,446 annotated CT images across four categories: cyst, tumor, stone, and normal. Experimental results demonstrate a substantial improvement over the baseline YOLOv8 model, achieving an overall classification accuracy of 91.47%, with precision, recall, and F1-scores of 90.32%, 88.76%, and 89.21%, respectively. Notably, the recall for the critical and often-missed tumor class improved from 30.41% to 82.34%. This work presents a significant step toward reliable, automated diagnostic support, offering a tool that can assist clinicians in early and accurate detection of renal pathologies, thereby improving patient outcomes.

Keywords : Kidney Abnormality Detection, Computed Tomography (CT), YOLOv8, Attention Mechanism.

Authors : S Mahammed Sajid , T Abdul Raheem , P Veeresh , Gandla Nagappa

Title : An Attention-Enhanced YOLOv8 Model For Accurate Multi-Class Kidney Abnormality Detection in CT Imaging

Volume/Issue : 2026;3(1 (January - March))

Page No : 68 - 73

Contrastive vision-language models (VLMs) like CLIP have shown that they can do well on zero-shot classification tasks, but their accuracy relies a lot on how effectively the prompts are built. To solve this problem, we use GPT-4 to make prompts that are more visually descriptive, which makes CLIP more flexible when working with fine-grained picture collections. We put up a whole pipeline on the CUB-200 bird species dataset, starting with preprocessing the images (resizing, converting to greyscale, and bilateral filtering) and then utilizing statistical measures and the Gray-Level Co-occurrence Matrix (GLCM) to extract features. The dataset is divided into training and testing sets, and VGG-19 is used for transfer learning to classify the data. This gives a solid visual baseline. To make the text space more interesting, GPT-4 is used to provide extensive visual explanations for each class label. These descriptions are then added to CLIP’s text embedding space. These embeddings match the outputs of CLIP’s visual encoder, which makes it easier to interpret things from different modes. Using criteria like accuracy, F1-score, precision, recall, and mistake rate to test the suggested technique shows that it works better than typical CLIP prompts for categorization. Finally, a web site is built using Python, Streamlit, HTML, and CSS that lets users input pictures and get predictions with full details. This study shows how feature-based preprocessing, deep transfer learning, and GPT-4-enhanced prompt engineering may work together to greatly improve zero-shot classification accuracy in fine-grained image recognition tasks.

Keywords : GPT-4, Classification, Transfer Learning, Computer Vision

Authors : Ramireddy Sasidhar Reddy , A V Santhosh Kumar

Title : A Hybrid Framework Combining Preprocessing, VGG-19 Transfer Learning, and GPT Enhanced CLIP for Fine-Grained Classification

Volume/Issue : 2026;3(1 (January - March))

Page No : 60 - 67

Reproducing machine learning results across teams and hardware often fails due to hidden randomness, drifting datasets, and unstable environments that alter metrics without visible changes to code or intent. ReproQuorum virtualizes randomness with scoped, counter‑based generators, derives hash‑locked master seeds from commit, dataset fingerprints, and canonicalized configuration, emits signed JSON‑LD provenance receipts, and packages a Nix‑pinned OCI capsule for deterministic re‑execution. Repository continuous integration targets a benchmark quorum of ≥ 2 public datasets per task; the reported experiments enforce ≥ 1 dataset per task. Across named entity recognition (WNUT 2017) and extractive question answering (SQuAD v1.1), replication attains Tier‑A on matched architectures (byte‑identical) and Tier‑B across CPU↔CUDA (metric‑identical with Δ≤10^(-6)). Performance improves by +1.1 F1 on WNUT and +0.7 F1 / +0.6 EM on SQuAD under equal budgets. Capsules and receipts enable one‑command verification and reduce ambiguity about seeds, data, and environments. Code, data manifests, checkpoints, capsules, and model cards are released with signed receipts and container digests to support independent validation.

Keywords : Reproducibility, Determinism, Provenance, Benchmark Quorum, OCI Containers, RNG Virtualization.

Authors : Surya Pavan Kumar Gudla , Preeti Nutipalli , T Chalapathi Rao

Title : ReproQuorum : Signed, Scope Deterministic Pipelines and Benchmark Quorums for Verifiable ML Reproducibility

Volume/Issue : 2026;3(1 (January - March))

Page No : 52 - 59

Stampedes pose a major risk to people's lives in densely populated settings, such as public gathering places, transit hubs, and places of worship. In order to avoid stampedes and fatalities, it is crucial to promptly recognize unstable crowds. In this paper, we offer a computer vision system that uses YOLOv8-based person detection to analyze crowd behavior and estimate the probability of a stampede. To ascertain how a crowd is acting collectively, the system makes use of data on time, motion patterns, and crowd density. Furthermore, we employed a spatial zone-based analytic technique to pinpoint regions of the scene with the highest potential risk of a stampede. Finally, we have included a way to build a temporal persistence on the population in order to reduce the incidence of false alarms and stabilize risk estimations over time. Through a real-time web-based interface, our technology generates visual analytics output and multi-level risk alerts. The suggested method can distinguish between crowds that are behaving normally, at the warning stage, and in a high-risk state, according to experimental results acquired from recorded videos of crowds. This creates a useful decision support tool for crowd monitoring applications.

Keywords : Computer Vision, Crowd Density, Motion Analysis, YOLOv8, Stampede Risk Detection, Crowd Analysis.

Authors : Kuppireddy Krishna Reddy , T Ramya , Yagireddi Ramesh , J V G Prakasa Rao Pyla

Title : Real-Time Stampede Risk Prediction from Crowd Videos Using YOLOv8 and Spatio-Temporal Modeling

Volume/Issue : 2026;3(1 (January - March))

Page No : 44 - 51

In this paper, an alternative technique for determining unexpected human-movement regions is presented and demonstrated. In particular, we propose an efficient alternative to human activity by bounding or identifying individual humans or their parts, denoting an actuator or effect -intensifying tracking. The proposed movement effect map serves as a comprehensive representation of motion velocity characteristics, which includes the movement direction and magnitude of objects or subjects that interact with each other inside a particular area or group. Record Terms: Strange movement discovery, imaginative and prescient-primarily based reconnaissance, movement impact map, swarmed scenes/domestic.

Keywords : Crowd Anomaly Detection, Spatio-Temporal Feature Learning, Video Surveillance Analytics, DL.

Authors : Kodeboina Suresh , Sk.Akbar , T Murali Mohan

Title : Spatio-Temporal Modeling for Abnormal Behaviour Detection in Crowd Scenes

Volume/Issue : 2026;3(1 (January - March))

Page No : 37 - 43

Cognitive Behavioural Therapy (CBT) has been extensively explored for its efficacy in addressing mental health challenges. The research demonstrates three applications which include treatment for paranoid schizophrenia with ongoing delusions and depression prediction through NLP Transformers that study social media content and mental health accessibility improvements through AI chatbot systems. The implementation of Virtual Reality technology together with Cognitive Behavioural Therapy enables university students to experience effective anxiety relief through its system. The development of ChatGLM-LoRA models for insomnia treatment demonstrates how CBT has evolved to meet contemporary needs of psychotherapeutic systems. The mental health treatment of various disorders shows the effective results which Cognitive Behavioural Therapy (CBT) delivers through its various technological advancements. Cognitive Behavioural Therapy (CBT) has been used to treat paranoid schizophrenia which is characterized by people experiencing continuous delusions and strong distrust towards others. The research showed that participants who were specifically chosen for the study achieved major progress in their understanding of delusions and their belief systems and behaviour patterns after they received therapy which proved to be effective at correcting their mental misperceptions. The chronic sleep disorder insomnia which affects sleep patterns and health has been treated through AI-driven CBT interventions that use ChatGLM-LoRA models to develop conversational applications which achieved clinical trial validation of their effectiveness. The virtual reality system found in CBT introduces a gamified experience, which enables university students to use 3D environments for learning CBT principles while they undergo anxiety treatment. The research demonstrates that CBTA has achieved wider availability through its use of AI technologies, which provide accessible, non-disruptive methods to treat complex mental health disorders.

Keywords : Talk bots, COVID-19, HMC, Social Media Emotional Support, Schizophrenia Prediction.

Authors : Ch Prathima , V N Chetan Kumar Pulipati , Kalla Vijay

Title : Leveraging ChatGLM-LoRA for Scalable Mental Health Interventions: A Cognitive Behavioural Therapy

Volume/Issue : 2026;3(1 (January - March))

Page No : 15 - 36

Identifying-matching the same individuals has become a complex thing in strengthening security and it shows seamless living within smart cities. This task involves detecting and matching a person across multiple camera views for surveillance and safety applications. The capability and ability are able to accurately recognize individuals from diversified visual data sources. The new pathways of computer vision accelerated the research progress with deep neural networks to process high quality surveillance data. By analyzing the structural elements of a Re-ID framework, one can differentiate between closed-world and open-world identification scenarios. This work proceeds shows data preprocessing strategies for person re-identification and evaluates deep learning approaches using widely recognized benchmark datasets. Finally, it mainly focus on the vital key role on developing an effective deep learning and that leverages various multimodal feature extraction aimed at performance analysis identification and performance accuracy. We proposed new technique evaluated to performance on human recognition capabilities.

Keywords : Person matching identification, Re-Identification, Hashing, CNN, Autoencoders, Artificially Damaged Frames.

Authors : T Ramya , Kuppireddy Krishna Reddy

Title : A Multimodal Deep Learning Framework for Robust Person Re-Identification

Volume/Issue : 2026;3(1 (January - March))

Page No : 8 - 13

Recently developed speech emotion recognition (SER) methods have placed much emphasis on the derivation of features of acoustic data. In this paper, an alternative method is described, which shows a higher accuracy of the SER implementation because it does not rely on any manual steps, such as feature extraction. The suggested approach is a mixture of Conformer blocks and CNNs which are used to predict the emotions directly based on the audio signals. The Conformer block is used to solicit long-range dependencies and time-specific features whereas CNN layers are good at extracting localized emotional information. In this design, detailed emotional content is maintained since it is otherwise overlooked due to the conventional approaches and at the same time the hierarchical depictions of emotional content are captured. The Conformer combination of time and situation data, and the CNN layers make special attention to extracting spatial features. The model was tested on three publicly available datasets and in more than one language and significantly improved accuracy and interpretability with the use of emotional and temporal information. This method moves toward affective computing since it offers a more detailed and effective analysis of the speech data.

Keywords : Affective Computing, Speech Emotion Recognition, Deep Learning, Conformer-CNN Hybrid, Interpretability.

Authors : S Guru Prasad , M Sreevani

Title : Enhancing Speech Emotion Recognition with Deep Learning Techniques

Volume/Issue : 2026;3(1 (January - March))

Page No : 1 - 7