BFORE: Optimized Retinex for Low-Light Image Enhancement
Enhance low-light images with BFORE, a hybrid Butterfly-Firefly optimized Retinex method delivering superior brightness and clarity without training data.
OpenAI Launches Trusted Contact to Prevent Self-Harm
OpenAI's new Trusted Contact feature on ChatGPT helps support users by alerting trusted individuals during potential self-harm situations.
MHPR Benchmark for Human Perception in Vision-Language AI
Discover MHPR, a new benchmark enhancing human perception and reasoning in large vision-language models for real-world AI applications.
MEMSAD: Advanced Anomaly Detection for Memory Poisoning
Discover MEMSAD, a novel defense against memory poisoning attacks in retrieval-augmented agents, enhancing AI security with gradient-coupled anomaly detect...
CuraView: AI Framework for Detecting Medical Hallucinations
CuraView uses GraphRAG and multi-agent AI to detect and verify hallucinations in medical records, improving accuracy and patient safety in EHRs.