A convolutional neural network (CNN) slides small filters across an image to detect features, building from edges up to objects. CNNs are compact and fast, which keeps them common on edge devices even as transformers lead on raw accuracy.
The book · Volume 6 of 7
AI for Video: A Complete Guide to Intelligence in Video Products: From Models and Inference to Computer Vision, Speech, Search, and Generative Video
A convolutional neural network is still the workhorse behind most detectors that have to run on a camera or a phone. This volume covers CNNs in depth as part of its computer vision section: the ML minimum for product decisions, detection, segmentation and pose models, model artifacts, and the deployment topologies that decide where a CNN runs. Written by Nikolay Sapunov, CEO at Fora Soft.
Choosing between a CNN and a transformer for your video model?
Tell us the task, the hardware and the frame rate, and we will come back with a model shortlist and a latency and accuracy estimate for each option. 250+ video products shipped since 2005.
