Abstract

Handwritten Text Recognition (HTR) presents a significant challenge in computer vision due to various factors such as individual writing styles, noise, blur, and other imperfections in the text. This challenge is further exacerbated when dealing with languages using Indian scripts, which are characterized by complex character structures, extensive character inventories, and specific cultural nuances. In this study, we address these challenges by focusing on enhancing handwritten text recognition for ten Indic languages: Hindi, Bengali, Telugu, Tamil, Gujarati, Gurumukhi, Oriya, Kannada, Malayalam, and Urdu. We aim to improve recognition accuracy by leveraging the Permuted Autoregressive Sequence Model (PARSeq), an extension of the transformer-based model. Our results demonstrate the superiority of the PARSeq model over existing approaches, particularly in achieving state-of-the-art performance across most languages. Additionally, we investigate the efficacy of transfer learning from printed text to handwritten text, revealing its potential to enhance recognition performance. The trained models and code are publicly available at https://github.com/LalithaEvani/Indic-HTR-CVIP-2024.

Abstract

Reliable toxic gas detection is vital for residential and industrial safety. While precise sensors are expensive, affordable ones face challenges of nonlinearity and environmental sensitivity, particularly from temperature (T) and humidity (H) effects. In this work, we present a novel predeployment calibration framework that accounts for these environmental factors on the sensor behavior, given by the resistance ratio (Rs/Ro) at constant gas concentration levels. The proposed method first refines the baseline resistance (Rs) estimation by fitting a power-law model to known gas concentrations and then applies a cubic regression model to capture the nonlinear effects of temperature and humidity on the Rs/Ro ratio. Cubic regression achieves superior accuracy (>5.8%) over lower order models while reducing over-fitting risks compared to higher order polynomials. It achieves 99.65% average accuracy, outperforming the 96.73% from standard libraries. This improved performance is particularly notable at low ppm levels, where direct Rs measurements are typically noisy and unstable. The enhanced stability and accuracy of the proposed method were validated over a continuous 90-min test period.

Abstract

Efficient algorithms for graph connected components and minimum spanning tree (MST) computations are vital in diverse domains, from medical diagnostics to chip design. Yet, as graph sizes grow to billions of edges, existing approaches often fail due to limited device memory. This challenge demands innovative rethinking of both algorithmic design and implementation strategies. We address this need by presenting a novel Parallel Heterogeneous External Memory (PHEM) framework, which harnesses both multicore CPU and GPU resources to tackle large-scale graph processing. Our method minimizes communication overhead while maximizing computational throughput through efficient copy- computation strategies. Experiments on a variety of real-world and synthetic benchmarks demonstrate that our algorithms for connected components and minimum spanning tree in the PHEM model outperform current state-of-the-art solutions, significantly advancing the capability to process massive graphs.

Abstract

Optical Character Recognition (OCR) for Indian languages is challenging due to diverse scripts, complex characters, and varied writing styles. This work presents HindiOCR-VLM, a Vision-Language Model adapted for Hindi OCR using Low-Rank Adaptation (LoRA) on a model pre-trained for Chinese and English. We explore two adaptation strategies: (i) a textit{single-stage model that directly predicts text at the page level --- bypassing intermediate word or line detection --- focused on printed documents, and (ii) a two-stage model for multi-domain scenarios (printed, handwritten, and scene text), which first detects words and then recognizes them individually. Inspired by human learning processes, we propose a progressive learning approach --- a training strategy to the textit{single-stage} model to enhance language acquisition and accelerate convergence. Leveraging the vision encoder’s rich representations, our method enables effective multi-domain training. Experimental results and ablation studies show that textit{HindiOCR-VLM} handles the complexities of the Devanagari script well and outperforms domain-specific models, offering a unified and robust solution for Hindi OCR. Code and resources are available at: https://hindiocr-vlm.github.io/.

Abstract

Globesity has widely affected the world population. It is a multifactorial health problem. Obesity can be caused by high-fat intake, high-calorie intake, physical inactivity, age, gender, or hormonal issues that can induce many health issues, such as cardiovascular disorders, diabetes, and metabolic disorders. Health facilities are insufficient to provide services to the whole population. In such a case, introducing Artificial Intelligence (AI) in the health sector is one of the most significant steps that can be taken. With the help of AI, the prognosis, diagnosis, and treatment can be provided to individuals within their homes and setups without any physical interactions. Assembling the data collected with AI can help to give more information on the type of obesity prevailing in a particular society so that it could be prevented in the population. Since AI learning is relatively novel within the existing system of medicine, enabling …

Abstract

Instruction fine-tuning of large language models (LLMs) is a powerful method for improving task-specific performance, but it can inadvertently lead to a phenomenon where models generate harmful responses when faced with malicious prompts. In this paper, we explore Low-Rank Adapter Fusion (LoRA) as a means to mitigate these risks while preserving the model’s ability to handle diverse instructions effectively. Through an extensive comparative analysis against established baselines using recognized benchmark datasets, we demonstrate a 42% reduction in the harmfulness rate by leveraging LoRA fusion between a task adapter and a safety adapter, the latter of which is specifically trained on our safety dataset. In addition, we made noteworthy observations related to exaggerated safety behavior, where the model rejects safe prompts that closely resemble unsafe ones.

Abstract

The partition of 1947 represents a significant rupture in the history of modern South Asia. Current literary research on partition primarily focuses on Hindi and Urdu narratives from India and Pakistan. This paper aims to study the use of silence in Bangladeshi partition stories and analyze the use of various narrative tools such as ellipses, deliberate pauses, omissions, and atmospheric cues by their authors for producing this phenomenon of silence. This paper claims that the use of silence in these Bangladeshi partition stories is strategic and not accidental. It also demonstrates that this narrative silence resembles the kind of silence present in the testimonies of partition survivors. This analysis of silence is based upon close readings of The Escape (1950) by Syed Waliullah, The Exile (1987) by Hasan Azizul Huq, and Roots (1998) by Syeda Farida Rahman. These stories feature themes that are representative of the experience of general people during partition like displacement, identity, belonging, memory, psychological trauma, communal and sexual violence, etc. These short stories are written by Bangladeshi authors from diverse sexual and social backgrounds, who also wrote across different periods spanning from the 1940s to the 1990s.

Abstract

This paper investigates the nuanced distinctions between typical and atypical speech disfluencies, focusing on features captured by self-supervised models. Typical disfluencies are natural, non-pathological irregularities in speech, while atypical disfluencies are linked to speech disorders like stuttering, characterized by more frequent and severe disruptions. Despite progress in automatic disfluency detection, limited research addresses the direct classification of these two types. This study leverages intermediate representations from four pretrained models Wav2Vec2.0, HuBERT, WavLM, and TERA, to analyze and classify typical and stuttered disfluencies. The experiments utilize two novel Indian English datasets, IIITH-IEDE and IIITH-TISA, enabling a comprehensive analysis of disfluency patterns in a linguistically diverse context. Classification experiments with support vector machines (SVM) and convolutional neural networks (CNN) reveal that features from HuBERT’s 5th layer—balancing low-level acoustic and high-level semantic information—achieve a peak F1 score of 0.97. These findings highlight the importance of intermediate layer representations of self-supervised models in distinguishing nuanced speech variations and contribute to robust and interpretable systems for speech disfluency classification.

Abstract

Recent advancements in multimodal large language models (MLLMs) have driven researchers to explore how well these models read data visualizations, e.g., bar charts, scatter plots. More recently, attention has shifted to visual question answering with maps (Map-VQA). However, Map-VQA research has primarily focused on choropleth maps, which cover only a limited range of thematic categories and visual analytical tasks. To address these gaps, we introduce MapIQ, a benchmark dataset comprising 14,706 question-answer pairs across three map types: choropleth maps, cartograms, and proportional symbol maps spanning topics from six distinct themes (e.g., housing, crime). We evaluate multiple MLLMs using six visual analytical tasks, comparing their performance against one another and a human baseline. An additional experiment examining the impact of map design changes (e.g., altered color schemes, modified legend designs, and removal of map elements) provides insights into the robustness and sensitivity of MLLMs, their reliance on internal geographic knowledge, and potential avenues for improving Map-VQA performance.

Abstract

Accurate and complete product descriptions are crucial for e-commerce, yet seller-provided information often falls short. Customer reviews offer valuable details but are laborious to sift through manually. We present PRAISE: Product Review Attribute Insight Structuring Engine, a novel system that uses Large Language Models (LLMs) to automatically extract, compare, and structure insights from customer reviews and seller descriptions. PRAISE provides users with an intuitive interface to identify missing, contradictory, or partially matching details between these two sources, presenting the discrepancies in a clear, structured format alongside supporting evidence from reviews. This allows sellers to easily enhance their product listings for clarity and persuasiveness, and buyers to better assess product reliability. Our demonstration showcases PRAISE's workflow, its effectiveness in generating actionable structured insights from unstructured reviews, and its potential to significantly improve the quality and trustworthiness of e-commerce product catalogs.