Learn about machine learning in image recognition

What is Machine Learning?

Machine learning gives computers the ability to learn without being explicitly programmed. Instead, it learns from annotated historical data. It powers key applications of 2017 including voice assistance, image understanding, and content recommendation.

What is Machine Learning?

Supervised learning

The goal of supervised learning is to learn patterns from historical data and find similar patterns in new samples. Input data must be annotated often by a human. Supervised learning is the most common machine learning and includes applications like image recognition, object detection and natural language processing.

Unsupervised learning

The goal of unsupervised learning is to find patterns in the set of unlabeled data. The common unsupervised method is cluster analysis and it is used to find hidden patterns or grouping in data. Its Application in computer vision is for example grouping images with similar features or style.

Machine learning

Machine learning neural network

Machine learning consists of simpler algorithms used in data analytics. Its benefit is good mathematical understanding of inner structure and performance. It contains algorithms like Linear Regression, Decision Tree or Random Forest.

Deep learning

Deep learning neural network

Deep learning uses many layers of neurons designed to solve a specific task like image recognition or face recognition
. It is very time and data consuming to train a deep network. Deep learning has developed in last few years thanks to the parallel processing using GPUs.

Machine learning on images?

General Tagging

General Tagging example

Living room, plant, sofa

Custom Tagging

Custom Tagging example

spacious, B&W, lamp, living room

Custom classification

General Tagging example

Accepted/not accepted; modern style

Segmentation

General Tagging example

Wall

Object Detection

Custom Tagging example

Home cinema (x,y,w,h)

Similarity

General Tagging example

40% similar

General recognition

  • Pretrained models (animal, food, face)
  • 1000s of quality tags/categories
  • High accuracy on general categories
  • Developed by Google, IBM, Microsoft

Custom image recognition

  • Initial training required
  • User-defined tags/categories
  • High accuracy on specific problems
  • Developed by Vize
Explore use cases Try it now

Machine learning production flow

Most of the machine learning applications are designed to improve in time based on a data feedback loop. Machine learning flow is based on new generated data and feedback for the current model performance. Model is periodically retrained to reflect the most relevant data.

Machine learning loop

Web platform feedback loop

Machine learning loop in cloud

Mobile platform feedback loop

Machine learning for mobile devices - Mobile Models - CoreML

API and Cloud

Vize Custom Image Recognition API

+ Feedback loop and improvements
+ Simple scaling and parallelism
+ Higher accuracy
+ Versioning
- One second latency
- Higher bandwidth


Vize Image Recognition cloud

On-device processing

Vize MobileModels

+ Low latency
+ Distributed performance
- Battery draining
- Lower accuracy
- No feedback loop


Apple - CoreML - machine learning framework Android - Tensorflow lite
Embedded - Nvidia Jetson DL kit

Technology Stack

Neural networks in numbers

Single image processing fires 8 billion floating point operations

Each model has 25 million of parameters

Neural network pre-trained on 50 million images

Each GPU provides 5 TFLOPS of processing power

Software

Hardware

Comparison between in-house development and SaaS

SaaS solution

+ Feedback loop and big data
+ Problem overview
+ ML optimization and meta-learning
+ Infrastructure and scaling
+ No development time and cost

×

In-house development

+ Covers specific scenarios
+ No data transfer
- High development time and costs

Get started freeContact sales

Free, easy to set up, no credit card required