Institute for Human Development

Recently, image super-resolution (ISR) has achieved notable advances facilitated by the advent of generative AI models relying on convolutional and adversarial architecture. However, an ongoing challenge is preserving fine structural detail, especially edges and textures, which can lead to blurry or over-smoothed images produced at high resolution. In this work, we introduce a novel generative framework entitled Edge-Aware Transformer for SuperResolution (EAT-SR), which utilizes hierarchical attention mechanisms and transformer-based architecture designed specifically for retaining edge fidelity. EAT-SR incorporates a Dynamic Edge Attention Module (DEAM) that learns to focus on important edge areas by dynamically learning attention weights based on the spatial and frequency domain. Additionally, we introduce a Multi-Scale Contextual Transformer (MCT) that captures global and local dependencies across multiple scales in order to improve texture fidelity while still maintaining coherence. Extensive evaluation on DIV2K, Set5, Set14, and Urban100 benchmark datasets demonstrates that EAT-SR outperforms other state-of-the-art methods by yielding improved PSNR (33.82 dB) and SSIM (0.937) scores, along with sharper edges and better texture restoration. The proposed method is also shown to be reliable in dimness and noise situations - for real time applications such as - medical imaging, satellite, and video enhancement. Overall, this study is an important step toward models for super-resolution that are perceptually accurate, computationally efficient in nature by integrating attention mechanism with generative AI frameworks and is believed to be a significant step forward in SR modeling.

Abstract In the East Asia-Pacific region, the Regional Comprehensive Economic Partnership ( RCEP ) agreement is expected to strengthen trade related to regional value chains ( RVC s) amongst RCEP members through reductions in tariff and non-tariff measures, implementation of liberal rules of origin (RoO), and trade cost reductions. This paper evaluates more comprehensive global value chain ( GVC ) indicators from the OECD Trade in Value Added (TiVA) database ( OECD 2023) and the Asian Development Bank (2022) dataset for RCEP countries and argues that the aggregate benefits from regional integration may mask the variety of asymmetric effects amongst participating countries. Despite specialisation in activities based on comparative advantage, the diversity amongst RCEP members could impact their potential share in total value-added created within GVC s and require state policies for stimulating economic transformation and economic upgrading.

IntroductionIt is predicted that low-and- middle-income countries like India will bear the maximum brunt of global aging in the near future. Aging societies are increasingly facing higher incidence of cognitive decline and dementia. Research on the potential of dietary patterns as a modifiable lifestyle factor for protection of cognitive health has gained momentum. This study aims at developing a dietary assessment tool-food frequency questionnaire (FFQ) for older Indian adults in the community.MethodologyAn FFQ was developed for the diverse, older population in India to collect data on their food intake over the past 12 months. This was pre and pilot tested leveraging the sample of an ongoing longitudinal study of late life cognition in older adults in India – Longitudinal Aging Study in India- Harmonized Diagnostic Assessment of Dementia (LASI-DAD). The FFQ was pilot tested during the first wave of the study (2017–2020) on a sample of 1125 respondents, updated and modified and finally implemented in the second wave of the study on 4465 respondents (2022–2024). An electronic Computer Assisted personal interview (CAPI) system with the use of tablets was implemented. The LASI-DAD FFQ was translated to 12 Indian languages for collecting accurate data from different regions of the country. A diverse range of food products and list of local food from different parts of India were included.ResultsThe challenges faced, lessons learnt and modifications helped shape the final LASI-DAD FFQ. This was implemented in 24 Indian states. It had 4 sections: (A) Food consumption habits (B) Food frequency section (C) Herbal /non herbal/ayurvedic supplements (D) Spice intake. The final questionnaire had 88 food items and median time taken to complete was 31 min. Response rate varied from 97 to 100%.Conclusion and discussionThe LASI-DAD FFQ proved to be a good tool for dietary data collection from a diverse, older Indian population. It is a single, comprehensive questionnaire adapted to the diverse dietary habits of older Indian population factoring in socio-economic, geographic, cultural, religious and seasonal variations, and it will help explore potential associations between diet and late-life cognition in older Indians.

Cross‐cultural adaptation of cognitive assessments is crucial for detecting Alzheimer's disease (AD) and related dementias (ADRD) in aging populations. This study documents the adaptation of the Harmonized Cognitive Assessment Protocol (HCAP) for a pilot study on the Longitudinal Study of Health and Aging in Kenya (LOSHAK) in rural Kilifi County, Kenya, highlighting challenges and strategies for optimizing outcomes. As part of the LOSHAK feasibility phase, cognitive tests including: the Swahili Mental State Examination, 10‐word recall, animal naming, story recall, clock drawing, and making change, were administered to 202 participants (≥45 years) from the Kaloleni/Rabai Health and Demographic Surveillance System (KRHDSS). Measures were adapted culturally and linguistically, and trained local enumerators conducted home‐based assessments. Low literacy (60.1% had no schooling), linguistic diversity, cultural norms, and infrastructure limitations influenced assessments. Key adaptations included translation, culturally relevant modifications, flexible administration, and community engagement. Contextualized cognitive assessments improve validity in rural resource‐limited settings, offering insights for future research.HighlightsThis is a narrative qualitative study documenting the experiences and Strategies Adopted to Optimize Effective Adaption of Cognitive Tests in Longitudinal Study of Health and Aging in Kenya (LOSHAK). This article documents the adaptation and contextualization of the Harmonized Cognitive Assessment Protocol (HCAP) for a pilot study on health and aging (LOSHAK) in rural Kilifi County, Kenya, focusing on the challenges and strategies employed to optimize outcome.HCAP tests were administered as part of the feasibility and pilot phase of the LOSHAK, the aim of which was to validate measures and optimize data collection procedures.The median age of the 202 study participants was 64, with 57.4% being female. A majority (62.5%) were not currently working, and nearly 70% fell within the two poorest wealth quintiles.The rural setting presented unique challenges including:low literacy rates (60.1% of participants had no schooling),diverse language use (primarily Giriama and Swahili),limited infrastructure (e.g., 44.8% of households had electricity),restrictive cultural norms that influenced data collection (e.g., in‐home interviews were often conducted outdoors with destructions).Key adaptations included:translating and culturally adapting test items (e.g., using local Swahili dialects and culturally relevant examples),recruiting and training local enumerators who were familiar with the community, culture and language,iterative pre‐testing and using roleplay helped to ensure that enumerator scoring was consistent, accurate, and reliable.contextualization of tool and data collection strategies included: adjusting data collection methods to accommodate cultural practices and environmental limitations, for example, allowing respondents to use their preferred language for “animal naming,” using paper and pen for some test items for which participants experienced difficulty using tablets, paying close special attention to time and season of administration, and adopting strategies to minimize background noise and other environmental distractions (e.g., due to rain, lunch hour, planting and harvesting, and school holidays). Prior to data collection, it was essential to engage local community health volunteers (CHVs), build rapport with participants, explain the study, and describe the idiosyncrasies of cognitive testing.The study emphasizes the importance of cultural sensitivity, linguistic appropriateness, and community engagement in cognitive assessment in diverse, resource‐constrained settings.The findings offer practical recommendations for researchers aiming to conduct cognitive assessments in similar populations, contributing to the development of culturally sensitive and effective tools for understanding and addressing cognitive health in aging populations.

Medical devices are one of the most crucial segment of a country’s healthcare system. This paper reviews the demand and supply side factors, including socio-economic, market-enabling, technological and policy instruments, that could drive the growth of the medical device industry. It highlights some key issues and challenges the sector is grappling with. Despite there being several enabling factors for the sector’s growth, India primarily manufactures medical equipment within the low-tech segment, from consumable to implantable devices. This leaves domestic requirements unmet in other segments, pushing the country to import expensive equipment in the advanced technology segment. This drives up the cost of medical equipment, leading to higher diagnostic test fees for end-users, which in turn places a significant out-of-pocket financial payment burden on households for diagnostic services.

Abstract This chapter studies inequality of opportunity in education in India and the key factors contributing these disparities. Using the ex-ante approach and machine learning algorithms, it analyses years of schooling as the outcome variable drawing from Periodic Labour Force Survey (PLFS) data. Input variables include gender, social group, parents’ education and occupation, and geographic region. The results reveal significant educational gap between rural and urban areas, as well as across gender and social groups. Gini coefficient of 0.22 indicate significant educational inequality, with about one-third of it stemming from circumstances beyond an individual’s control. Among all, parental education and geographical location emerge as the primary factors contributing to educational inequality. The study highlight the need for targeted policies to improve educational schools in underserved region, support marginalized communities, and promote gender equality.