A York University study in Current Biology used the motion aftereffect, the waterfall illusion, to show that people and macaque IT cortex shift where they see an object after watching motion, while AI vision networks from VGG-16 to SlowFast do not. What it means and what it does not show.
Chiba University’s Hiroyuki Kubo uses ordinary projectors and cameras to reveal veins beneath the skin and the direction of flowing water, and builds AI that proposes colours for anime artists to choose. We explain the computational imaging behind the Q&A, trace the papers since 2018 and set out the lessons for businesses using computer vision.
Computer scientist Bo Wang explains why AI that shines in the lab stumbles in homes, hospitals and streets. We walk through his three challenges with the numbers from his own research, from pose models that collapse in low light to vision-language models that spot unseen actions, plus a readiness checklist.
How AI Sees the City: Urban Visual Intelligence,” a new Routledge book from the MIT Senseable City Lab and Peking University, argues that visual AI can turn the images cities already produce into detailed data on traffic, emissions, greenery and public space. MIT research in New York used 331 traffic cameras to show congestion pricing cut traffic by about 10% but emissions by 16% to 22%. The authors also warn of surveillance and bias. This article explains the method, the findings, the risks and the questions to ask before deploying visual AI in public spaces.
A study by Mugihiko Kato and Biyu J. He of New York University, published in iScience on 17 September 2026, built an image set that systematically untangles global shape, internal parts and texture, then compared human viewers against more than 200 deep neural networks. No model reproduced the human cue-reliance profile, and every model tested substantially underperformed people when global shape was the only usable cue. This article covers the design, the finding that brain-alignment scores did not predict behavioural alignment, which model families came closest, and what the result means for autonomous driving, robotics and assistive devices.
The Wall Street Journal reports that OpenAI bought Glass Imaging, the start-up behind the GlassAI neural camera processor used in Honor’s 600 series, for more than $300 million. We set out what is confirmed, who founded the company, how its technology works and why OpenAI may want a camera team.
Lawrence Livermore’s camera-based inspection system measures 3D-printed strands layer by layer. We read the npj Advanced Manufacturing paper behind it: 13,797 labelled images, a 3.40 µm error that becomes 7.48 µm against the identity line, a 25 cm cushion that exposed a tilted build plate, and a 100,000x speed claim the paper never states.
A 14-page paper from the School of Art at East China Jiaotong University became the headline “AI-powered VR helps planners design better cities”. We read its 225-word abstract and the publisher’s 215-word research note. The abstract reports four measured results, including a 0.034-metre median geometric error. The note quotes none of them, and neither text says any planner used the tool.
MS-COCO is the benchmark almost every object detector reports against, and a new Sejong University survey has just published a table that flags it for all seven annotation error types it defines. The same survey publishes a second table that counts errors in six benchmarks by hand. MS-COCO is not in it. That combination — […]
Annotation errors are not a rounding error in the benchmarks the computer vision industry has trusted for a decade. Bounding boxes and class labels are the ground truth that every artificial intelligence vision system is trained on and then graded against, so a mistake in the labels is a mistake the model inherits twice over. […]
Progressive Robot: Your Gateway to Comprehensive IT Solutions — Specializing in Web Development, Mobile App Development, and Expert IT Services.
© All Copyright by Progressiverobot.com
VAT Number ( 506152326 )
