CrowdBA: A Low-Cost Quality-Driven Crowdsourcing Architecture for Bounding Box Annotation Based on Blockchain

Electronics, Journal Year: 2025, Volume and Issue: 14(2), P. 345 - 345

Published: Jan. 17, 2025

Many blockchain-based crowdsourcing frameworks currently struggle to address the high costs associated with on-chain storage and computation effectively, they lack a quality-driven incentive mechanism tailored bounding box annotation scenarios. To these challenges, this paper proposes CrowdBA: A low-cost, architecture. The CrowdBA utilizes Ethereum public blockchain as foundational architecture develops corresponding smart contracts. First, by integrating InterPlanetary File System (IPFS), processes are shifted off-chain, effectively addressing data on blockchains. Additionally, introduces Dynamic Intersection over union-weighted fusion (DWBF) algorithm, which assigns dynamic weights based IoU infer true boxes, thereby assessing each worker’s quality. Annotation quality then serves key criterion for distribution, ensuring fair appropriate compensation all contributors. Experimental results demonstrate that operational of contract function remain within reasonable limits; off-chain approach significantly reduces expenses, DWBF algorithm shows marked improvements in accuracy robustness other methods.

Language: Английский

IoT, Blockchain, Big Data and Artificial Intelligence (IBBA) Framework—For Real-Time Food Safety Monitoring

Applied Sciences, Journal Year: 2024, Volume and Issue: 15(1), P. 105 - 105

Published: Dec. 26, 2024

Technological advancements in mechanized food production have expanded markets beyond geographical boundaries. At the same time, risk of contamination has increased severalfold, often resulting significant damage terms wastage, economic loss to producers, danger public health, or all these. In general, governments across world recognized importance having safety processes place impose recalls as required. However, primary challenges existing practices are delays identifying unsafe food, siloed data handling, delayed decision making, and tracing source contamination. Leveraging Internet Things (IoT), 5G, blockchains, cloud computing, big data, a novel framework been proposed address current challenges. The enables real-time gathering situ application machine learning-powered algorithms predict facilitate instant making. Since processed real approach be identified early informed decisions made confidently, thereby helping reduce significantly. also throws up new implementation changes collection phases production, onboarding various stockholders, adaptation process.

Language: Английский