Evaluation of Winning Solutions of 2025 Low Power Computer Vision Challenge

2026-04-21

AI Summary

This work addresses the challenge of deploying efficient vision models on resource-constrained edge devices, where accuracy must be balanced against latency, memory footprint, and energy consumption. It presents a systematic evaluation of winning solutions from the three tracks of the 2025 Low-Power Computer Vision Challenge—robust image classification under varying illumination and style, open-vocabulary segmentation guided by text prompts, and monocular depth estimation—using a unified, reproducible edge evaluation framework built on Qualcomm AI Hub for the first time. By integrating model compression, hardware-aware network design, and multimodal prompting mechanisms, the study synthesizes emerging trends in efficient vision model development and competition-driven optimization strategies. The results demonstrate the feasibility of achieving high accuracy under stringent edge constraints and establish a methodological benchmark for future research and challenges in low-power computer vision.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Learning on the Edge & Model Compression
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Security and privacy of machine learning and AI applications

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.

Problem

• Low-Power Computer Vision
• Edge Devices
• Efficient Vision Models
• Open-Vocabulary Segmentation
• Monocular Depth Estimation

Innovation

• Low-Power Computer Vision
• Edge AI
• Qualcomm AI Hub
• Open-Vocabulary Segmentation
• Monocular Depth Estimation