Computer vision - Category Home
Computer vision - Topics
Why generative video models still struggle with anything longer than a few seconds
Turing-style tests for AI images and music
Tracking by detection and how identity switches accumulate
Semantic vs instance vs panoptic segmentation
Skip connections: why ResNets can train hundreds of layers
How ImageNet reshaped what counted as progress in computer vision
Why vision models collapse under unfamiliar lighting
Adding video and robotics to the test: the total Turing test
SLAM in rooms full of moving people
Optical flow: Lucas-Kanade vs learned dense flow
Industrial defect detection when the faulty class has fifty images
Testing vision models on your own camera instead of public benchmarks
Construction site safety monitoring with computer vision
Fusing face and voice signals for emotion classification
Grasping unfamiliar objects is still the hardest robotics problem
CNN receptive fields: how stacked 3x3 convolutions beat one 7x7
IoU and mAP: how object detection is actually scored
Depth sensing: stereo vs structured light vs time-of-flight
OCR on scanned documents with skewed and faded text
Data augmentation for vision: cutout and mixup do more than flipping
Camera calibration and lens distortion quietly ruin measurements
Camera modules and depth sensors that work with open-source vision models
3D reconstruction with structure from motion and neural radiance fields
Can facial expression recognition really tell what someone feels?