Evaluation of Winning Solutions of 2025 Low Power Computer Vision Challenge

arXiv cs.CV / 4/22/2026

📰 NewsDeveloper Stack & InfrastructureSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The IEEE Low-Power Computer Vision Challenge (LPCVC) focuses on building efficient vision models for edge devices by balancing accuracy against latency, memory, and energy constraints.
The 2025 edition ran three tracks—image classification under varying conditions, open-vocabulary text-prompt segmentation, and monocular depth estimation—to test different low-power vision capabilities.
The paper describes the LPCVC 2025 competition design and evaluation framework, emphasizing reproducible benchmarking via the Qualcomm AI Hub.
It also highlights the best solutions for each track and summarizes emerging trends and observations from the competition.
Finally, the authors provide recommendations for how future computer vision challenges could be structured and evaluated.

Abstract

The IEEE Low-Power Computer Vision Challenge (LPCVC) aims to promote the development of efficient vision models for edge devices, balancing accuracy with constraints such as latency, memory capacity, and energy use. The 2025 challenge featured three tracks: (1) Image classification under various lighting conditions and styles, (2) Open-Vocabulary Segmentation with Text Prompt, and (3) Monocular Depth Estimation. This paper presents the design of LPCVC 2025, including its competition structure and evaluation framework, which integrates the Qualcomm AI Hub for consistent and reproducible benchmarking. The paper also introduces the top-performing solutions from each track and outlines key trends and observations. The paper concludes with suggestions for future computer vision competitions.