About

D.Waste is an AI-powered waste management platform that empowers individuals and communities to make smarter waste disposal decisions. Our app uses AI to identify types of waste, guides users on proper disposal, and rewards sustainable behavior through a gamified points system.

D.Waste isn't just for individuals — we partner with educational institutions, organizations, and government bodies to promote large-scale behavioral change and sustainability:

• Educational Institutions: D.Waste helps schools and universities integrate environmental learning into the classroom. Our tools include real-time waste identification, sustainability tracking dashboards, and interactive educational games such as DSort that make learning about waste sorting fun and engaging for kids.
• Organizations: We support workplaces and nonprofits in running green initiatives by offering smart bin tagging, staff engagement through our reward system, and monthly sustainability reports to measure impact.
• Government Bodies: D.Waste provides municipalities and local governments with the tools to raise awareness, reduce landfill dependency, and monitor community behavior. Our analytics dashboard delivers valuable data for designing better waste policies and public outreach campaigns.

We also build interactive web-based games like DSort to teach young users how to sort waste through play. These games include drag-and-drop mechanics and puzzles that reward learning with points and badges. They are accessible on all devices and can be integrated into classroom or community outreach programs.

Available on the App Store and Google Play Store, D.Waste also helps users locate recycling centers, access educational materials, and track their personal impact over time.

Partner with us — Whether you're a school, NGO, or city council, D.Waste is ready to collaborate and help you build a cleaner, smarter future. Get in touch to start your sustainability journey with us.

Our technology is backed by peer-reviewed research and has been cited by many researchers worldwide. Over the years, we have published multiple papers, presented at international workshops and summits, and released open datasets that advance the global understanding of waste management and recycling.

Poster

DWaste: Greener AI for Waste Sorting using Mobile and Edge Devices

Oct 2025

AI + Environment Summit

The rise of the convenience packaging has led to generation of enormous waste. Sorting these waste is crucial in waste management. To aid the sorting process we provide an AI powered platform call DWaste that runs on smartphones and edge devices, even without internet access. We trained image classification (EfficientNetV2S/M, ResNet50/101) and light weighted object detection (YOLOv8n, YOLOv11n) models using a subset of our own Waste dataset and annotated it using the custom tool Annotated Lab. Among these models, EfficientNetV2S achieved 96% accuracy but took longer inference time (0.22s) and higher carbon emission, whereas the lightweight object detection delivered 76–77% mAP, with low emissions, ultra-fast inference (~0.03s) with smaller model sizes(<7MB), making them ideal for real-time use on low-power devices. Quantization further reduced model size and VRAM usage by nearly half. Our experiment show how greener AI model support real-time waste sorting to manage waste.

View →

Poster

AI-Powered Material Detection for Solid Waste Management

Jul 2025

Artificial Intelligence for Materials Science (AIMS) Workshop

DWaste uses lightweighted and greener AI to detect, sort and track waste using mobile and edge devices -even offline. This poster presents an overview of the end-to-end workflow of DWaste, from data collection and annotation (using our custom tool, Annotate Lab) to model training and evaluation. We benchmark a subset of our dataset (Garbage Dataset) with several classification models (EfficientNetV2S/M, MobileNet, ResNet50/101) and detection models (YOLOv8n/v11n), with accuracy, carbon emissions (using CodeCarbon), and training time. Results show the EfficientNet achieved higher accuracy (~95%); YOLO models offered significantly lower emissions and faster performance, making them ideal for real-time, energy-efficient deployment. DWaste shows the use of greener AI to sort and manage waste.

View →

Paper

Health Care Waste Classification Using Deep Learning Aligned with Nepal's Bin Color Guidelines

Aug 2025

arXiv Preprint

The increasing number of Health Care facilities in Nepal has also added up the challenges on managing health care waste (HCW). Improper segregation and disposal of HCW leads to the contamination, spreading of infectious diseases and puts a risk of waste handlers. This study benchmarks the state of the art waste classification models: ResNeXt-50, EfficientNet-B0, MobileNetV3-S, YOLOv8-n and YOLOv5-s using Stratified K-fold techniques where we use 5 folds on combined HCW data, and found that the YOLOv5-s achieved higher of 95.06% accuracy but fell short few milliseconds in inference speed with YOLOv8-n model. The EfficientNet-B0 showed promising results of 93.22% accuracy but took the highest inference time. A repetitive ANOVA was performed to see statistical significance and the best performing model (YOLOv5-s) was deployed to the web with mapped bin color using Nepal's HCW management standards for public usage. Further work on the data was suggested along with localized context.

View →

Paper

Plastic Waste Detection Using Deep Learning: Insights from the WaDaBa Dataset

Mar 2025

Industrial and Domestic Waste Management

With the increasing use of plastic, the challenges associated with managing plastic waste have become more difficult, emphasizing the need for effective classification and recycling solutions. This study explored the potential of deep learning, focusing on convolutional neural networks (CNNs) and object detection models like YOLO to tackle this issue using the WaDaBa dataset. The results indicated that YOLO-11m achieved the highest accuracy (98.03%) and mAP50 (0.990), while YOLO-11n performed similarly but achieved the highest mAP50 (0.992). Lightweight models like YOLO-10n trained faster but had lower accuracy, whereas MobileNetV2 demonstrated impressive performance (97.12% accuracy) but fell short in object detection. YOLO-11n had the fastest inference time (0.2720s), making it ideal for real-time detection, while YOLO-10m was the slowest (5.9416s). Among CNNs, ResNet50 had the best inference time (1.3260s), whereas MobileNetV2 was the slowest (1.4991s). These findings suggested that by balancing accuracy and computational efficiency, these models could contribute to scalable waste management solutions. The study recommended increasing the dataset size for better generalization, enhancing augmentation techniques, and developing real-time solutions.

View →

Paper

Deep Learning in Waste Management: A Brief Survey

Jan 2025

International Journal of Complexity in Applied Science and Technology

The rapid growth of the global population is causing a significant increase in waste production, leading to serious environmental and public health challenges. To address these issues, waste management systems are incorporating advanced technologies. Machine learning and computer vision are used to predict waste patterns, optimise collection schedules, and improve sorting accuracy. Deep learning automates the sorting process, provides predictive analytics, and enhances recycling rates. Robotics, combined with AI and computer vision, improves sorting efficiency, while the internet of things (IoT) monitors waste levels and optimises collection routes. Despite these benefits, challenges such as data scarcity, high computational demands, and the need for substantial infrastructure investments must be addressed. This research explores the integration of advanced technologies into waste management and evaluates their effectiveness using waste datasets. It highlights the potential to tackle environmental challenges and lay the groundwork for more intelligent waste management solutions.

View →

Research software

Annotate-Lab: Simplifying Image Annotation

Nov 2024

The Journal of Open Source Software

Annotate-Lab is an open-source image annotation tool designed to enhance the efficiency and accuracy of the annotation process through a client-server architecture. The application consists of a React-based client that offers an intuitive user interface for performing annotations, and a Flask-based server that manages the storage and generation of annotated images and configuration settings. This robust combination makes Annotate-Lab a versatile solution for various fields requiring precise image labeling, such as machine learning, medical imaging, and autonomous driving.

View →

Dataset

Garbage Dataset

Dec 2023

Kaggle

This dataset contains images of garbage items categorized into 10 classes, designed for machine learning and computer vision projects focusing on recycling and waste management. It is ideal for building classification or object detection models or developing AI-powered solutions for sustainable waste disposal.

View →

Paper

Managing Household Waste Through Transfer Learning

Mar 2024

Industrial and Domestic Waste Management

As the world continues to face the challenges of climate change, it is crucial to consider the environmental impact of the technologies we use. In this study, we investigate the performance and computational carbon emissions of various transfer learning models for garbage classification. We examine the MobileNet, ResNet50, ResNet101, and EfficientNetV2S and EfficientNetV2M models. Our findings indicate that the EfficientNetV2 family achieves the highest accuracy, recall, f1-score, and IoU values. However, the EfficientNetV2M model requires more time and produces higher carbon emissions. ResNet50 outperforms ResNet110 in terms of accuracy, recall, f1-score, and IoU, but it has a larger carbon footprint. We conclude that EfficientNetV2S is the most sustainable and accurate model with 96.41% accuracy. Our research highlights the significance of considering the ecological impact of machine learning models in garbage classification.

View →

Paper

MWaste: An app that uses deep learning to manage household waste

Jul 2023

Clean Technologies and Recycling

The rapid growth of the global population is causing a significant increase in waste production, leading to serious environmental and public health challenges. To address these issues, waste management systems are incorporating advanced technologies. Machine learning and computer vision are used to predict waste patterns, optimise collection schedules, and improve sorting accuracy. Deep learning automates the sorting process, provides predictive analytics, and enhances recycling rates. Robotics, combined with AI and computer vision, improves sorting efficiency, while the internet of things (IoT) monitors waste levels and optimises collection routes. Despite these benefits, challenges such as data scarcity, high computational demands, and the need for substantial infrastructure investments must be addressed. This research explores the integration of advanced technologies into waste management and evaluates their effectiveness using waste datasets. It highlights the potential to tackle environmental challenges and lay the groundwork for more intelligent waste management solutions.

View →

Vision & Values