Image datasets released by Google to speed machine learning

Share this on social media:

Google’s Open Images and YouTube8-M have released datasets of annotated images and video to aid researchers develop new image analysis techniques.

The tagged Open Images dataset has 9 million entries, while YouTube8-M’s database contains 8 million videos with 50,000 hours of footage.

The datasets have been made available to further the development of machine learning algorithms, a technique whereby a machine can learn to recognise content in images based on tagged data previously supplied to it. Machine learning potentially offers more accurate image analysis software, but requires large volumes of data to do so.

Halcon 13, the latest version of MVTec’s machine vision library, uses deep learning algorithms, notably for OCR which, according to the company speaking at the Vision show in Stuttgart in November, gives a read rate two times faster than earlier OCR tools, but requires around 50 million images to cover all possible characters.

Most industrial vision software companies are experimenting with machine learning techniques in one way or another. Swiss company Vidi Systems is one early industrial image analysis software library that uses deep learning algorithms – it was shortlisted for the Vision Award at Vision 2016.

The Open Images and YouTube8-M datasets could also be useful for engineers developing embedded vision solutions.

Related articles:

Image processing reaches new depths - Facebook, Amazon and Google are all working on high-profile deep learning projects, from speech pattern recognition to building driverless cars. Rob Ashwell looks at how the technology is being deployed in the machine vision sector to improve and speed inspection

Recent News

25 May 2021

The face recognition imager consumes 10,000 times less energy than a typical camera and processor. CEA-Leti is working with STMicroelectronics on the imager

06 May 2021

The GTOF0503 sensor features a 5µm three-tap iToF pixel, incorporating an array with a resolution of 640 x 480 pixels

30 April 2021

The algorithm can deduce the shape, size and layout of a room by measuring the time it takes for sound from speakers to return to the phone's microphone

20 April 2021

The Kria K26 SOM is built on top of the Zynq UltraScale+ MPSoC architecture. It has 4GB of DDR4 memory and 245 IOs for connecting sensors