computer vision

A collection of 248 posts

3D semantic segmentation

Unlocking Smarter 3D Vision: How Geometric AI Biases Enhance Scene Understanding

Explore GIBLy, a lightweight AI layer that dramatically improves 3D semantic segmentation by embedding geometric priors, making AI more accurate and efficient for real-world applications.

AI hallucinations

Auditing Autonomous AI: Unpacking Trajectory-Level Hallucinations in Industrial Workflows

Explore trajectory-level hallucinations in multi-agent AI systems for industrial workflows. Learn about a new taxonomy and evaluation framework crucial for reliable enterprise AI deployment.

AI video surveillance

AI-Powered Surveillance: An Interpretable Framework for Suicide Risk Assessment in Metro Stations

Explore an innovative AI framework using video analytics for proactive suicide prevention in public spaces. Learn how AI assesses risk through behavioral patterns, spatial context, and temporal dynamics.

AI in medical imaging

Beyond Simulation: AI Diffusion Models Create Hyper-Realistic PET Images for Medical Advancement

Explore how a novel AI diffusion model, PAD, synthesizes highly realistic, heterogeneous PET images from simple organ maps. This breakthrough revolutionizes medical research, AI training, and virtual trials by overcoming traditional simulation limits. Learn about its precision and real-world applica

Industrial anomaly detection

JUDO: Pioneering Domain-Oriented AI for Industrial Anomaly Detection

Discover JUDO, a novel AI framework enhancing Large Multimodal Models with internalized domain knowledge for superior industrial anomaly detection, explainability, and reliability. Learn how it transforms manufacturing and operations.

multimodal AI

Combating Misinformation: How Multimodal AI Detects Fake News with Fuzzy Logic

Explore F2IND-IT!, a novel multimodal AI framework leveraging fuzzy logic, ResNet-50, and DistilBERT to detect fake news by analyzing both images and text, with a focus on its critical role in countries like India.

Medical AI

Bridging the Learnability Gap: Advancing AI for Rare Medical Conditions

Discover the "learnability gap" in medical latent diffusion models and how innovative AI techniques are enhancing diagnostic accuracy for rare diseases. Explore practical applications and ARSA's role in delivering robust healthcare AI solutions.

Thermal crowd counting

Thermal-Only Crowd Counting: Advancing Privacy and Accuracy with AI

Explore ARSA Technology's insights into thermal-only crowd counting, leveraging AI and diffusion models to enhance accuracy while eliminating RGB data for superior privacy in surveillance.

Data augmentation

Beyond Accuracy: How Data Augmentation Reshapes AI's Internal Understanding

Explore how data augmentation strategies geometrically transform neural network representations, offering new insights for optimizing AI performance and developing robust enterprise solutions.

IoT sensing

Dywave: Revolutionizing IoT Sensing with Event-Aligned Dynamic Tokenization

Discover Dywave, an AI framework that transforms heterogeneous IoT signals into compact, event-aligned tokens. Enhance accuracy and efficiency for real-time analytics in smart cities, healthcare, and industrial IoT.

Spiking neural networks

Advancing Edge AI: Breaking Through SNN Transformer Bottlenecks for Practical Vision Applications

Explore how Local Structure-Aware Spiking Transformers (LSFormers) overcome computational and data loss challenges in SNNs, enabling highly accurate, energy-efficient AI for real-world vision applications.

Transformer Spiking Elastis

Transformer Spiking Elastis: Solusi AI Hemat Energi untuk Pemahaman Gerakan di Perangkat Edge

Temukan Elastic Spiking Transformers (NESTformer) dari penelitian terbaru, arsitektur AI adaptif yang merevolusi pemahaman gerakan di perangkat edge. Hemat energi hingga 60% tanpa mengurangi akurasi.

Dense 3D Tracking

TrackCraft3R: Repurposing Video Diffusion Transformers for Cutting-Edge Dense 3D Tracking

Discover TrackCraft3R, a pioneering method that leverages video diffusion transformers for dense 3D tracking, offering unmatched speed, accuracy, and efficiency for dynamic scene analysis.

Low-light image enhancement

Multi-Modal AI: Revolutionizing Low-Light Image Enhancement with M2Retinexformer

Explore M2Retinexformer, a pioneering multi-modal AI framework enhancing low-light images using depth, luminance, and semantic cues. Improve visibility, reduce noise, and boost AI performance in critical applications.

AI air purifier

AI-Powered Air Purifiers: How Edge AI and Computer Vision Deliver Personalized Air Quality with Enhanced Privacy

Explore how advanced air purifiers use on-device AI and computer vision to deliver personalized air quality, track users, and ensure data privacy through edge processing. Discover the broader implications for smart environments.

Edge Deep Learning

Driving Innovation: The Power of Edge Deep Learning in Computer Vision and Medical Diagnostics

Explore how Edge Deep Learning transforms computer vision and medical diagnostics. Learn about optimized AI on edge devices, real-time applications, and future trends. Discover ARSA's practical AI solutions.

AI Authorship Attribution

Unmasking Online Criminal Networks: How AI Authorship Attribution Links Digital Trafficking

Explore how advanced AI and machine learning, particularly authorship attribution, help law enforcement identify and link online criminal activities, including human trafficking, by analyzing digital stylistic patterns in text and images.

Explainable AI

Unlocking Trust: A User-Centric Look at Explainable AI in Medical Image Diagnosis

Explore how Explainable AI (XAI) transforms medical image diagnosis. Discover physician preferences for visual and textual explanations and the crucial role of user-centric design for AI adoption in healthcare.

Plant Disease Detection

Revolutionizing Agriculture: Lightweight AI for Rapid Plant Disease Detection

Discover how lightweight Multi-View Convolutional Neural Networks (MV-CNNs) are transforming plant disease identification, offering fast, accurate, and cost-effective solutions for sustainable agriculture, even in remote areas.

Explainable AI

Unlocking AI Transparency: How Max-Plus Neural Networks Deliver Explainable Decisions

Explore max-plus neural networks for explainable AI. Learn how their unique architecture provides transparent decision-making, offering crucial insights for critical applications like healthcare and industry.

Image Dehazing

AI-Powered Image Dehazing: Revolutionizing Clarity in Non-Homogeneous Environments

Discover how multi-branch AI networks are overcoming the challenges of non-homogeneous haze, delivering unprecedented image clarity for critical enterprise applications like autonomous driving and surveillance.

AI video analytics

Revolusi Data Sintetis: Diagnosis Kegagalan Model Visi AI dengan Eksperimen Terdesain

Pelajari bagaimana Synthetic Designed Experiments (SDRS) merevolusi diagnosis kegagalan model visi AI, menghemat biaya, dan meningkatkan akurasi dengan data sintetis yang ditargetkan secara efisien.

Hyperspectral Image Classification

MixerCA: Advancing Hyperspectral Image Classification with Efficient AI

Explore MixerCA, a novel AI model leveraging depthwise convolution and coordinate attention for efficient, high-accuracy hyperspectral image classification. Discover its practical applications in remote sensing.

Energy-efficient AI

Energy-First AI: Revolutionizing Neural Networks for Sustainable Performance

Explore how energy-first neural network design, inspired by biological principles, optimizes AI for both accuracy and efficiency, leading to sustainable enterprise solutions.

Edge AI optimization

Optimizing Edge AI: A Comparative Analysis of Advanced UCB Algorithms in Adaptive Deep Neural Networks

Explore how advanced Upper Confidence Bound (UCB) algorithms enhance Adaptive Deep Neural Networks (ADNNs) for efficient, low-latency AI at the edge. Discover insights into accuracy, energy, and latency trade-offs for enterprise deployments.