No abstract

Because of the rise of cryptocurrencies and decentralized apps, blockchain technology has generated a lot of interest. Among these is the emergent blockchain-based crowdsourcing paradigm, which eliminates the centralized conventional mechanism servers in favor of smart contracts for task and reward allocation. However, there are a few crucial challenges that must be resolved properly. For starters, most reputation-based systems favor high-performing employees. Secondly, the crowdsourcing platform’s expensive service charges may obstruct the growth of crowdsourcing. Finally, unequal evaluation and reward allocation might lead to job dissatisfaction. As a result, the aforementioned issues will substantially impede the development of blockchain-based crowdsourcing systems. In this study, we introduce ExCrowd, a blockchain-based crowdsourcing system that employs a smart contract as a trustworthy authority to properly select workers, assess inputs, and award incentives while maintaining user privacy. Exploration-based crowdsourcing employs the hyperbolic learning curve model based on the conduct of workers and analyzes worker performance patterns using a decision tree technique. We specifically present the architecture of our framework, on which we establish a concrete scheme. Using a real-world dataset, we implement our model on the Ethereum public test network leveraging its reliability, adaptability, scalability, and rich statefulness. The results of our experiments demonstrate the efficiency, usefulness, and adaptability of our proposed system.

No abstract

Human pose estimation has long been a fundamental problem in computer vision and artificial intelligence. Prominent among the 2D human pose estimation (HPE) methods are the regression-based approaches, which have been proven to achieve excellent results. However, the ground-truth labels are usually inherently ambiguous in challenging cases such as motion blur, occlusions, and truncation, leading to poor performance measurement and lower levels of accuracy. In this paper, we propose Cofopose, which is a two-stage approach consisting of a person and keypoint detection transformers for 2D human pose estimation. Cofopose is composed of conditional cross-attention, a conditional DEtection TRansformer (conditional DETR), and an encoder-decoder in the transformer framework; this allows it to achieve person and keypoint detection. In a significant departure from other approaches, we use conditional cross-attention and fine-tune conditional DETR for our person detection, and encoder-decoders in the transformers for our keypoint detection. Cofopose was extensively evaluated using two benchmark datasets, MS COCO and MPII, achieving an improved performance with significant margins over the existing state-of-the-art frameworks.

No abstract