Resource-constrained Environments Research Articles

Transforming Agriculture through Artificial Intelligence: Advancements in Plant Disease Detection, Applications, and Challenges

Artificial intelligence (AI) has emerged as a revolutionary tool in agriculture, particularly in the realm of plant disease detection. This article provides an overview of AI-powered plant disease detection methods, their applications, and the limitations associated with their implementation. By leveraging AI, farmers can enhance crop management practices, optimize resource utilization, and mitigate yield losses caused by plant diseases. However, challenges such as data scarcity, model interpretability, and deployment in resource-constrained environments remain significant barriers to widespread adoption. Addressing these limitations is crucial for maximizing the potential of AI in revolutionizing agriculture and ensuring global food security. Artificial intelligence (AI) has emerged as a pivotal tool in modernizing agriculture, particularly in the domain of plant disease detection. This article presents a comprehensive examination of AI-driven methodologies for plant disease detection, exploring their applications and inherent limitations. Through the utilization of machine learning and computer vision techniques, AI facilitates early disease identification, precision agriculture, disease surveillance, and decision support systems. Despite these transformative capabilities, challenges such as inadequate data availability, model interpretability, and implementation in resource-constrained settings impede widespread adoption. Addressing these obstacles is imperative for fully harnessing the potential of AI in agricultural innovation, thereby safeguarding global food security and sustainability.

Cardiovascular disease risk assessment through sensing the circulating microbiome with perovskite quantum dots leveraging deep learning models for bacterial species selection.

Perovskite quantum dots (PQDs) are novel nanomaterials wherein perovskites are used to formulate quantum dots (QDs). The present study utilizes the excellent fluorescence quantum yields of these nanomaterials to detect 16S rRNA of circulating microbiome for risk assessment of cardiovascular diseases (CVDs). A long short-term memory (LSTM) deep learning model was used to find the association of the circulating bacterial species with CVD risk, which showed the abundance of three different bacterial species (Bauldia litoralis (BL), Hymenobacter properus (HYM), and Virgisporangium myanmarense (VIG)). The observations suggested that the developed nano-sensor provides high sensitivity, selectivity, and applicability. The observed sensitivities for Bauldia litoralis, Hymenobacter properus, and Virgisporangium myanmarense were 0.606, 0.300, and 0.281fg, respectively. The developed sensor eliminates the need for labelling, amplification, quantification, and biochemical assessments, which are more labour-intensive, time-consuming, and less reliable. Due to the rapid detection time, user-friendly nature, and stability, the proposed method has a significant advantage in facilitating point-of-care testing of CVDs in the future. This may also facilitate easy integration of the approach into various healthcare settings, making it accessible and valuable for resource-constrained environments.

Integrating technology into a successful apomorphine delivery program in Thailand: a 10-year journey of achievements with a five-motto concept.

Apomorphine, a potent dopamine agonist, is a therapeutic option for patients with Parkinson's disease and motor fluctuations. However, the adoption of and adherence to this therapy have been limited by the need for complex delivery devices and specialized care as well as resource consumption, posing challenges for new physicians. Thailand is a unique example of a developing nation that has successfully implemented and continued the use of this therapy by employing cooperative technology that has dramatically enhanced apomorphine delivery services. Establishing apomorphine delivery services requires significant resources and step-by-step solutions. We began our services by implementing various strategies in three chronological stages: the initial stage (2013-2015), intermediate stage (2016-2019), and current stage (2020-present), each presenting unique challenges. Together, we also implemented a proposed set of five mottos to strengthen our apomorphine delivery service. Using additive technology, we developed a patient registry platform that combined electronic data acquisition, video and remote monitoring using wearable sensors, and in-house mobile applications to support our service. At the initial stage, we assembled a team to enhance the efficacy and confirm the safety of apomorphine treatment in our hospital. At the intermediate stage, we expanded our apomorphine delivery services beyond just the patients at our hospital. We supported other hospitals in Thailand in setting up their own apomorphine services by educating both physicians and nurses regarding apomorphine therapy. With this educational undertaking, increased apomorphine-related knowledge among medical professionals, and a greater number of hospitals providing apomorphine services, an increasing number of patients were administered apomorphine in subsequent years. Currently, we are providing effective apomorphine delivery to improve patient outcomes and are seamlessly integrating technology into clinical practice. Incorporating integrative technologies in our apomorphine delivery program yielded positive results in data collection and support throughout patient care, in tracking patients' statuses, in the long-term use of this treatment, and in increasing medication adherence rates. This perspective paper describes how technology can help provide supportive healthcare services in resource-constrained environments, such as in Thailand, offering a step-by-step approach to overcoming several limitations. The valuable insights from our 10-year journey in successfully integrating technology into apomorphine delivery services can benefit new physicians seeking to replicate our success.

Detection of dental periapical lesions using retinex based image enhancement and lightweight deep learning model

Dental periapical lesions, commonly associated with inflammation around the tooth apex, pose a significant challenge in early diagnosis and treatment. This study introduces a novel approach for the detection of dental periapical lesions through the integration of Retinex-based image enhancement techniques and a lightweight deep learning model. The Retinex algorithm is employed to enhance the radiographic images, addressing issues related to inconsistent illumination and contrast. Subsequently, a tailored lightweight deep learning model is designed to efficiently extract relevant features from the enhanced images. Present methodology leverages a dataset of dental radiographs to train and evaluate the deep learning model, incorporating a diverse range of periapical lesion cases. The model is optimized for computational efficiency while maintaining high accuracy, making it suitable for deployment in resource-constrained environment. To enhance precision in lesion detection, the U-Net segmentation technique has been incorporated, providing a sophisticated approach to delineate and analyze specific areas of interest within the radiographic images. This addition further refines our diagnostic framework, contributing to the robustness of lesion identification. Experimental results demonstrate the effectiveness of the Retinex-based image enhancement in improving the visibility of periapical lesions. The lightweight deep learning model exhibits promising performance in accurately detecting and classifying dental periapical lesions, showcasing its potential for early and efficient diagnosis. The results obtained from the present model are compared with those from Convolutional Neural Network (CNN) as well as with diagnosis of expert practitioners and the model is observed to perform very well. The study contributes to the advancement of computer-aided diagnostic tools in dentistry, offering a scalable and accessible solution for the identification of dental periapical lesions through the fusion of image enhancement and lightweight deep learning techniques.

AI Enabled MED Drone for Healthcare Application

The project proposes the development of an AI-enabled MedDrone for medication delivery, introducing an innovative and efficient solution to enhance healthcare logistics. The MedDrone, equipped with artificial intelligence capabilities, aims to revolutionize the process of medication delivery by leveraging autonomous aerial technology. The system integrates advanced algorithms to optimize route planning, ensuring timely and accurate delivery of medications to designated locations. The AI component enables the drone to adapt to real-time variables such as weather conditions and traffic, enhancing the reliability and responsiveness of the delivery process. This project addresses the challenges of traditional medication distribution, offering a futuristic and intelligent approach to healthcare logistics through the deployment of AI-enabled autonomous drones for efficient and timely medication delivery. The proposed system integrates AI algorithms for real-time route optimization, obstacle avoidance, and decision-making, enabling the drone to autonomously navigate complex environments and deliver essential medical supplies such as first aid kits, defibrillators, and medications to the point of need. Through the utilization of machine learning techniques, the drone continuously learns from its interactions with the environment, enhancing its adaptability and performance over time. Additionally, the system incorporates advanced communication technologies to enable seamless coordination with emergency responders and healthcare professionals, ensuring timely delivery and proper utilization of resources. The efficacy of the proposed AI-based medical drone system is demonstrated through simulations and real-world trials, highlighting its potential to revolutionize emergency response efforts and improve healthcare accessibility, particularly in challenging or resource-constrained environments.

Comprehensive Survey of Model Compression and Speed up for Vision Transformers

Vision Transformers (ViT) have marked a paradigm shift in computer vision, outperforming state-of-the-art models across diverse tasks. However, their practical deployment is hampered by high computational and memory demands. This study addresses the challenge by evaluating four primary model compression techniques: quantization, low-rank approximation, knowledge distillation, and pruning. We methodically analyze and compare the efficacy of these techniques and their combinations in optimizing ViTs for resource-constrained environments. Our comprehensive experimental evaluation demonstrates that these methods facilitate a balanced compromise between model accuracy and computational efficiency, paving the way for wider application in edge computing devices.

A rehearsal framework for computational efficiency in online continual learning

In the realm of online continual learning, models are expected to adapt to an ever-changing environment. One of the most persistent hurdles in this adaptation is the mitigation of a phenomenon called "Catastrophic Forgetting" (CF). This critical condition occurs when models trained on non-identically distributed data lose performance in previously learned tasks. Rehearsal methods, leveraging the ability to replay older samples, aim to address this challenge by incorporating a buffer of past training samples. However, the absence of known task boundaries complicates the adaptation of current CF mitigation methods. This paper proposes a method attuned to data stream characteristics and online model performance in a resource-constrained environment. The number of training iterations and learning rate emerges as crucial hyperparameters, impacting the efficacy and efficiency of online continual learning. Up to this point, we propose a combination of Experience Replay methodologies, a Drift Detector, and various training convergence policies, specially tailored for scenarios with unknown task boundaries. Experimental results demonstrate the effectiveness of our approach, maintaining or enhancing performance compared to baseline methods, while significantly improving computational efficiency.

81 A rapid-cycle application of the Consolidated Framework for Implementation Research allows timely identification of barriers and facilitators to implementing the World Health Organization’s Emergency Care Toolkit in Zambia

OBJECTIVES/GOALS: Implementation science evaluations are often too time-intensive to provide actionable feedback during implementation, suggesting the need for more agile methods. We present an evaluation of the World Health Organization’s Emergency Care Toolkit implementation in Zambia using rapid qualitative methods to provide timely feedback. METHODS/STUDY POPULATION: We evaluated the implementation of the Emergency Care Toolkit in eight general and referral hospitals in Zambia in 2023 using a rapid-cycle, qualitative template analysis approach grounded in the Consolidated Framework for Implementation Research (CFIR). We gathered qualitative data from operational field notes, focus groups, and key informant interviews of administrators, clinicians, nurses, and support staff in all eight hospitals in Zambia. We parsimoniously applied CFIR constructs and tool-specific codes, focused on barriers and facilitators, to allow for rapid but comprehensive cross-case analysis. The results were used to generate a matrix of stakeholder-relevant, plain-language barriers and facilitators for each tool. RESULTS/ANTICIPATED RESULTS: We completed eight site visits with focus groups and interviews following initial implementation in September 2023 to gather firsthand knowledge related to implementation of the Toolkit. The CFIR-focused coding accelerated analysis by centering on barriers and facilitators for each tool while maintaining a comprehensive evaluation framework. Summary tables of barriers and facilitators were easily interpreted by lay stakeholders. Visualization in tables allowed for identification of common themes across tools and hospitals, making comprehensive recommendations to the implementation and dissemination process quickly possible. We anticipate the study findings will empower implementing partners to make timely, actionable improvements. DISCUSSION/SIGNIFICANCE: Rapid-cycle qualitative implementation evaluations allow for rigorous yet timely feedback on the implementation process compared to traditional methods. This efficient strategy is particularly important in resource-constrained environments where inefficient implementation wastes limited resources and create delays that cost lives.

Utilization of Antenatal Care Services Within the Context of COVID-19, Security Challenges, and an Unstable Healthcare System at Primary Health Care Centers.

Utilizing antenatal care (ANC) is vital for maternal and neonatal well-being, especially in low-resource settings with healthcare challenges. Identifying factors impacting antenatal care this context, t of current situation in Sudan is crucial for developing strategies to improve maternal care. This study explores ANC utilization amidst COVID-19, security issues, and healthcare instability in primary health care centers in a resource-constrained environment, aiming to enhance maternal care access and quality. This is a mixed-methods study. Quantitative data were collected through a retrospective analysis of ANC attendance records before and during the pandemic. Demographic variables were analyzed for their association with ANC utilization. In-depth interviews were conducted to collect qualitative data from pregnant women and healthcare providers. These interviews focused on capturing the experiences, perceptions, and obstacles associated with antenatal care services during the pandemic and within the healthcare system challenges. Preliminary quantitative analysis revealed a significant decline in ANC utilization during the pandemic. Fear of COVID-19 infection, disrupted transportation, and resource constraints have emerged as significant barriers to ANC attendance. The qualitative findings highlighted the impact of security challenges on healthcare access and the adaptation strategies employed, including mobile ANC clinics and telehealth consultations. This study illuminates the intricate landscape of ANC utilization in the face of a pandemic, security instability, and healthcare system vulnerability. Policymakers and stakeholders should collaborate to strengthen healthcare systems and ensure the continued provision of essential maternal health services during challenging times.

Award-Winning Women Business Operators in Kandy District of Sri Lanka: Entrepreneurial Bricolage

Women entrepreneurship has been identified as a crucial factor in closing the gender gap in the labor market and women empowerment. However, there is a significant gender gap in the SME sector in Sri Lanka. Also, Literature supports that Sri Lankan women are living in a socially and economically constrained environment. In this study entrepreneurial bricolage has been used as a theoretical lens to understand the entrepreneurial actions of women business operators because bricolage links means and ends in a resource-constrained environment. Since women business operators in Sri Lanka are in a constrained environment and operating with limited resources, the main purpose of this study to understand how Award-winning women business operators have applied Bricolage approach in their entrepreneurship process. Qualitative multiple case study research design was used to achieve the research objectives. In-depth interview method was adopted as the main data collection method and 30 women awardees were selected by using the purposive sampling method. The findings illustrate that these women business operators have used the 'Bricolage' approach in the form of 'selective bricolage'. They have used it as a one-off practice to find solutions for problems and opportunities they faced in their businesses.

Enhancing environmental policy through evidence synthesis: a review of the Environmental Evidence for the Future (EEF) Initiative

The Environmental Evidence for the Future (EEF) Initiative emerged in response to the challenges and opportunities presented by the UK’s decision to leave the European Union and its associated Environmental Frameworks. The Natural Environment Research Council (NERC), working closely with the Collaboration for Environmental Evidence (CEE) and UK stakeholders, developed the initiative to identify and address crucial evidence gaps, offering a long-term vision for environmental policy and sustainability. The EEF Initiative progressed through three stages: strategic priority identification, NERC panel award selection, and the production of Systematic Maps of existing evidence. The first stage involved collaborative workshops across the UK to identify key knowledge gaps in environmental science. The subsequent prioritisation resulted in 65 challenges across 10 thematic areas. The second stage saw NERC initiating, with CEE support, an open call for research proposals emphasising the use of evidence synthesis methodology. The selection process, balancing topic importance and applicant expertise, led to funding for five projects. The final stage involved the production of Systematic Maps of existing evidence based on the CEE Guidelines and Standards, providing a structured overview of existing literature on specific topics. The EEF Initiative demonstrated effective collaboration between UKRI (NERC), an independent non-profit (CEE), academia, and government agencies, addressing critical environmental challenges through rigorous evidence synthesis methodologies. The programme enhanced understanding and utilisation of these methodologies within the research community. Key lessons include the importance of inclusive priority-setting, differentiation between broad policy questions and specific Systematic Map questions, recognition of the value of Systematic Maps, and the role of experience in evidence synthesis teams. As policymakers and researchers navigate environmental policies in a resource-constrained environment, the EEF Initiative highlights the cost-effectiveness and efficiency of systematic mapping and review processes for evidence-based decision-making. The success of funding through NERC sets a precedent for future thematic evidence focused programmes, emphasising the need for continued support in developing synthesis skills among researchers and encouraging direct government commissions for targeted and responsive evidence. The EEF Initiative serves as a model for effective collaboration, providing valuable insights into addressing evidence gaps and shaping evidence-based policymaking in an ever-evolving environmental landscape.

Principles, applications, and advancements of the Segment Anything Model

The Segment Anything Model (SAM) is a prominent computer vision model discussed in a review paper focusing on image segmentation. This paper explores the concepts, applications, and advancements of SAM, which excels at accurately separating diverse object types and managing visual data. It leverages convolutional neural networks (CNNs), an encoder-decoder architecture, skip connections, and spatial attention mechanism to capture fine details and contextual information across different scales. SAM finds versatile applications in various domains, including medical imaging for precise anatomical structure delineation and pathology identification. It improves recognition and classification by precise positioning and segmentation. However, the SAM model faces challenges such as complex object shapes and computational requirements for real-time deployment in resource-constrained environments. To tackle these limitations, researchers have proposed advancements like feature enhancement, network architecture modifications, and regularization techniques. Future directions may involve lightweight network designs, optimization strategies, and integration of external information to enhance accuracy, efficiency, and robustness of the SAM model.

Road Infrastructure, Supply Chain Costs Reduction, and Road Safety in Economic Geography Perspective

Purpose: While it is expected that road transport infrastructure may enhance supply chain cost reduction, the mechanisms and conditions under which road transport infrastructure may enhance supply chain cost reduction are less understood. Theoretical framework: Drawing insights from the economic geography perspective, and survey data from 359 senior managers of manufacturing firms in Ghana, we found support for the argument that road transport infrastructure is important in the quest to improve supply chain cost reduction. Methods: The hypothesized model was tested with survey data from beverage manufacturing firms in Ghana. The survey method is usually appropriate for strategic management studies. Results and Conclusion: Our findings also suggest that road safety does not just support supply chain cost reduction but serves as a transformative mechanism to reap superior supply chain cost reduction in a resource-constrained environment. Our findings make a contemporary contribution to the economic geography theory as well as important managerial guidance for managers in the supply chain management space in resource-constrained regions like SSA. Research implications: The outcome of this study offers an important extension to the frontiers of innovative literature in supply chain management. Originality/value: It also provides practical insight to practitioners in the manufacturing setting by emphasizing the need to pay crucial attention to road safety as well as invest in road infrastructure development in the quest to reduce supply chain costs.

Techniques to alleviate blockchain bloat: Potentials, challenges, and recommendations

Blockchain (BC) bloat, characterized by excessive data growth on BC networks, presents a pressing challenge. The need for a comprehensive study addressing BC bloat solutions arises due to the scarcity of knowledge in this emerging area. This study advances knowledge along with practical insights and inspires innovative approaches to address BC bloat. This research explores various techniques proposed to combat BC bloat. Two technique themes are identified: on-chain and off-chain techniques. On-chain techniques focus on updating the consensus mechanism, block size, or database partition. This study identified five on-chain techniques: structural modification, pruning, sharding, ephemeral BC, and zk-SNARKs. Off-chain techniques focus on processing in parallel, bundling transactions, or handling transactions off-chain. Further, three off-chain techniques identified are off-chain storage, historical data storage, and light clients’ techniques. These recommendations aim to optimize resources use, enhance scalability, and maintain data integrity. However, these may compromise several BC capabilities, such as security, privacy, and data loss. The applicability of these techniques for resource-constrained devices is also explored. Future research should include further investigation into the performance and applicability of these techniques, especially in resource-constrained environments like IoT and fog computing.

Using EEG total energy as a noninvasively tracking of intracranial (and cerebral perfussion) pressure in an animal model: A pilot study

Purpose: This study aims to describe the total EEG energy during episodes of intracranial hypertension (IH) and evaluate its potential as a classification feature for IH. New methods: We computed the sample correlation coefficient between intracranial pressure (ICP) and the total EEG energy. Additionally, a generalized additive model was employed to assess the relationship between arterial blood pressure (ABP), total EEG energy, and the odds of IH. Results: The median sample cross-correlation between total EEG energy and ICP was 0.7, and for cerebral perfusion pressure (CPP) was 0.55. Moreover, the proposed model exhibited an accuracy of 0.70, sensitivity of 0.53, specificity of 0.79, precision of 0.54, F1-score of 0.54, and an AUC of 0.7. Comparison with existing methods: The only existing comparable methods, up to our knowledge, use 13 variables as predictor of IH, our model uses only 3, our model, as it is an extension of the generalized model is interpretable and it achieves the same performance. Conclusion: These findings hold promise for the advancement of multimodal monitoring systems in neurocritical care and the development of a non-invasive ICP monitoring tool, particularly in resource-constrained environments.

Exploring sparsity in graph transformers

Graph Transformers (GTs) have achieved impressive results on various graph-related tasks. However, the huge computational cost of GTs hinders their deployment and application, especially in resource-constrained environments. Therefore, in this paper, we explore the feasibility of sparsifying GTs, a significant yet under-explored topic. We first discuss the redundancy of GTs based on the characteristics of existing GT models, and then propose a comprehensive Graph Transformer SParsification (GTSP) framework that helps to reduce the computational complexity of GTs from four dimensions: the input graph data, attention heads, model layers, and model weights. Specifically, GTSP designs differentiable masks for each individual compressible component, enabling effective end-to-end pruning. We examine our GTSP through extensive experiments on prominent GTs, including GraphTrans, Graphormer, and GraphGPS. The experimental results demonstrate that GTSP effectively reduces computational costs, with only marginal decreases in accuracy or, in some instances, even improvements. For example, GTSP results in a 30% reduction in Floating Point Operations while contributing to a 1.8% increase in Area Under the Curve accuracy on the OGBG-HIV dataset. Furthermore, we provide several insights on the characteristics of attention heads and the behavior of attention mechanisms, all of which have immense potential to inspire future research endeavors in this domain. Our code is available at https://github.com/LiuChuang0059/GTSP.

Analysing the impact on groundwater quality using dynamic programming and vision transformer

The demand for sustainable water sources has intensified the focus on ground water harvesting, necessitating rigorous quality analysis for diverse applications. The existing methods for ground water quality analysis are briefly discussed, revealing a research gap in optimizing the sampling and testing process. To address this gap, we employ Dynamic Programming, which is utilized to streamline the sampling and testing process by selecting optimal locations and times. This study proposes a new methodology that combines Dynamic Programming (DP) and the Vision Transformer (ViT) model to evaluate the impact on ground water quality efficiently. This optimization not only enhances data comprehensiveness but also minimizes resource utilization, a crucial aspect in resource-constrained environments. The incorporation of the Vision Transformer model into the methodology brings a novel dimension to ground water quality analysis. ViT, a cutting-edge deep learning model for image analysis, is adapted to process visual data from ground water samples. This includes identifying contaminants, suspended solids, and microbial content, providing a more holistic understanding of ground water quality than traditional methods. Results from our integrated approach demonstrate its efficiency and accuracy in assessing ground water quality. It achieves a 10–15% boost in sampling optimization efficiency, 5–10% improvement in visual data accuracy, and 5–10% increase in precision. DP-ViT reduces resource utilization by 15–20% and consistently offers 5–10% higher data coverage.

Some Like It Small: Czech Semantic Embedding Models for Industry Applications

This article focuses on the development and evaluation of Small-sized Czech sentence embedding models. Small models are important components for real-time industry applications in resource-constrained environments. Given the limited availability of labeled Czech data, alternative approaches, including pre-training, knowledge distillation, and unsupervised contrastive fine-tuning, are investigated. Comprehensive intrinsic and extrinsic analyses are conducted, showcasing the competitive performance of our models compared to significantly larger counterparts, with approximately 8 times smaller size and 5 times faster speed than conventional Base-sized models. To promote cooperation and reproducibility, both the models and the evaluation pipeline are made publicly accessible. Ultimately, this article presents practical applications of the developed sentence embedding models in Seznam.cz, the Czech search engine. These models have effectively replaced previous counterparts, enhancing the overall search experience for instance, in organic search, featured snippets, and image search. This transition has yielded improved performance.

MIND: Multi-Task Incremental Network Distillation

The recent surge of pervasive devices that generate dynamic data streams has underscored the necessity for learning systems to adapt continually to data distributional shifts. To tackle this challenge, the research community has put forth a spectrum of methodologies, including the demanding pursuit of class-incremental learning without replay data. In this study, we present MIND, a parameter isolation method that aims to significantly enhance the performance of replay-free solutions and achieve state-of-the-art results on several widely studied datasets. Our approach introduces two main contributions: two alternative distillation procedures that significantly improve the efficiency of MIND increasing the accumulated knowledge of each sub-network, and the optimization of the BachNorm layers across tasks inside the sub-networks. Overall, MIND outperforms all the state-of-the-art methods for rehearsal-free Class-Incremental learning (with an increment in classification accuracy of approx. +6% on CIFAR-100/10 and +10% on TinyImageNet/10) reaching up to approx. +40% accuracy in Domain-Incremental scenarios. Moreover, we ablated each contribution to demonstrate its impact on performance improvement. Our results showcase the superior performance of MIND indicating its potential for addressing the challenges posed by Class-incremental and Domain-Incremental learning in resource-constrained environments.

Depth Sensing Imaging System Autonomous Restored Fog

Edge-preserving smoothing is an image processing technique that smooths away textures while retaining sharp edges. Image de-noising is the technique to reduce noises from corrupted images. The aim of the image denoising is to improve the contrast of the image or perception of information in images for human viewers or to provide better output for other automated image processing techniques. Outdoor images taken in hazy climate often get degraded due to the effect of haze. There are several methods to remove haze from hazy images. Most of them over saturates the dehazed images. This degrades the quality of images. The color attenuation prior technique is one of the best algorithm to remove haze from images. Based on this technique and by making use of MATLAB software, this paper suggests a simple method to remove haze from Images. The core of DSIS lies in its fusion of depth information with traditional imaging data. By leveraging depth maps obtained from sensors such as LiDAR or structured light cameras, DSIS accurately delineates objects in the scene and their respective depths. This enables selective fog removal, prioritizing objects closer to the camera for clearer representation while preserving depth cues for improved scene understanding. Our project aims to implement and evaluate DSIS in real-world scenarios, such as outdoor surveillance and automotive vision systems. Through comprehensive testing and performance analysis, we seek to demonstrate the effectiveness of DSIS in restoring visibility under varying fog densities and lighting conditions. Additionally, we will explore optimization techniques to enhance DSIS's computational efficiency, ensuring its practical feasibility for deployment in resource-constrained environments.