I AM LISTENING TO
|
WHAT I LIKE
  • English
  • German


BLOG FILTER



Artificial intelligence (AI) has revolutionized the way we interact with images, and the current AI image solutions are a testament to this. AI image solutions are applications of AI that can identify, classify, and manipulate images with remarkable accuracy and speed. With the rise of deep learning and computer vision, AI image solutions have become increasingly sophisticated, and their applications have expanded to fields such as healthcare, finance, entertainment, and more. Some of the most prominent AI image solutions and their applications.

Image Recognition

Image recognition is the process of identifying and classifying objects, people, or other elements in an image. AI-powered image recognition solutions use deep learning algorithms to recognize images with high accuracy. Image recognition is widely used in various fields, including healthcare, security, and retail. For example, in healthcare, AI image solutions are used to analyze medical images such as X-rays and MRIs to detect diseases and conditions such as cancer, brain injuries, and more. In security, AI-powered image recognition solutions are used to identify faces, license plates, and other elements in surveillance videos. In retail, AI image solutions are used to identify and classify products and improve inventory management.

Object Detection

Object detection is a subset of image recognition that involves detecting the location of specific objects in an image. AI-powered object detection solutions can identify and locate objects in an image with high accuracy. Object detection is used in various fields, including self-driving cars, security, and e-commerce. For example, in self-driving cars, AI image solutions are used to detect pedestrians, traffic lights, and other objects on the road. In security, AI-powered object detection solutions are used to identify suspicious behavior and detect objects such as weapons and explosives. In e-commerce, AI image solutions are used to detect and locate products in images, improving search and recommendation algorithms.

  1. TensorFlow: TensorFlow is an open-source platform for building and training machine learning models, including image recognition models. It provides a wide range of tools and resources for image processing and can be used for tasks such as image classification, object detection, and segmentation.
  2. Keras: Keras is a high-level neural networks API written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It is a user-friendly and efficient way to build and train deep learning models, including image recognition models.
  3. PyTorch: PyTorch is another popular open-source machine learning framework that can be used for image recognition. It provides a flexible and easy-to-use platform for building and training deep learning models and supports various image processing tasks.
  4. OpenCV: OpenCV is a widely used computer vision library that provides tools and algorithms for image processing, including image recognition. It can be used for tasks such as object detection, face recognition, and motion detection.
  5. YOLO: YOLO (You Only Look Once) is an open-source object detection system that can recognize objects in images and videos with high accuracy and speed. It is based on deep learning and can be trained on custom datasets.
  6. ImageAI: ImageAI is an open-source library that provides pre-trained models for image recognition tasks such as object detection, classification, and segmentation. It supports several deep learning frameworks, including TensorFlow, Keras, and PyTorch.
  7. Detectron2: Detectron2 is an open-source object detection framework built on top of PyTorch. It provides a flexible and efficient platform for building and training object detection models and supports various architectures such as Faster R-CNN, Mask R-CNN, and RetinaNet.
  8. MXNet GluonCV: MXNet GluonCV is an open-source computer vision library built on top of MXNet. It provides a wide range of tools and resources for building and training object detection models and supports various architectures such as Faster R-CNN, Mask R-CNN, and SSD.
  9. SimpleDet: SimpleDet is an open-source object detection framework based on PyTorch. It provides a simple and efficient platform for building and training object detection models and supports various architectures such as Faster R-CNN and RetinaNet.

Image Segmentation

Image segmentation is the process of dividing an image into multiple segments or regions based on specific criteria. AI-powered image segmentation solutions use deep learning algorithms to identify different objects and elements in an image and segment them accordingly. Image segmentation is widely used in various fields, including healthcare, entertainment, and transportation. For example, in healthcare, AI image solutions are used to segment medical images such as CT scans and MRIs to aid in diagnosis and treatment. In entertainment, AI-powered image segmentation solutions are used to create special effects and manipulate images in movies and games. In transportation, AI image solutions are used to segment images of roads and traffic to aid in autonomous driving.

  1. Mask R-CNN: Mask R-CNN is an open-source framework based on deep learning that can perform instance segmentation, which involves identifying and segmenting each object in an image. It is built on top of the Faster R-CNN architecture and supports various deep learning frameworks such as TensorFlow and PyTorch.
  2. U-Net: U-Net is a popular open-source deep learning architecture for image segmentation. It is based on a fully convolutional neural network and has been widely used for tasks such as medical image segmentation.
  3. DeepLab: DeepLab is an open-source image segmentation framework based on deep learning. It supports various architectures such as DeepLab v3 and DeepLab v3+, which have achieved state-of-the-art performance on several benchmarks.
  4. SegNet: SegNet is an open-source deep learning architecture for image segmentation. It is based on an encoder-decoder network and has been widely used for tasks such as road segmentation and object detection.
  5. PyTorch-UNet: PyTorch-UNet is an open-source implementation of the U-Net architecture in PyTorch. It provides a simple and efficient platform for building and training image segmentation models.
  6. SimpleITK: SimpleITK is an open-source library for image analysis that provides tools and algorithms for image segmentation, registration, and filtering. It supports various programming languages such as Python, C++, and Java.

Image Generation

Image generation is the process of creating new images using AI algorithms. AI-powered image generation solutions can generate highly realistic images based on specific inputs, such as text descriptions or reference images. Image generation is used in various fields, including art, fashion, and advertising. For example, in art, AI image solutions are used to generate unique and creative designs and artworks. In fashion, AI-powered image generation solutions are used to create new designs and prototypes. In advertising, AI image solutions are used to generate highly realistic product images and visualizations.

  1. GAN (Generative Adversarial Networks): GANs are a type of deep learning model that can generate realistic images by training two neural networks, a generator and a discriminator, in a adversarial setting. Several open-source implementations of GANs are available, including TensorFlow-GAN, PyTorch-GAN, and Keras-GAN.
  2. StyleGAN (Style-Based Generative Adversarial Networks): StyleGAN is a type of GAN that can generate high-quality images with fine-grained control over the image style and content. It has been used to generate realistic human faces, animals, and other objects. An open-source implementation of StyleGAN is available on GitHub.
  3. DeepDream: DeepDream is a visualization technique that uses convolutional neural networks to generate abstract and surreal images. It works by optimizing the input image to maximize the activation of specific neurons in the network. An open-source implementation of DeepDream is available in the TensorFlow library.
  4. PixelRNN (Pixel Recurrent Neural Networks): PixelRNN is a type of neural network that can generate images pixel by pixel. It works by predicting the probability distribution of each pixel value based on the context of previous pixels. An open-source implementation of PixelRNN is available in the PyTorch library.
  5. VAE (Variational Autoencoder): VAEs are a type of deep learning model that can generate images by learning a low-dimensional representation of the image space. They work by encoding an input image into a lower-dimensional latent space and then decoding it back into an image. Several open-source implementations of VAEs are available, including TensorFlow-VAE and PyTorch-VAE.

AI image solutions are transforming the way we interact with images and are being used in various fields to improve efficiency, accuracy, and creativity. From image recognition and object detection to image segmentation and generation, AI image solutions are increasingly sophisticated and capable. As AI technology continues to evolve, we can expect to see more applications of AI image solutions in the future, and they will undoubtedly play an essential role in many industries.

The Future

Some of the tools are already used in new ways, to help reconstruct or understand archaeology dig sites. They help analyse genes and will help to discover new remedies or cures for illnesses in the future.

Even though many complain about the exponential growth of AI, it brings so many positive angels into the mix, allowing us to fix and elevate our lives. Sadly change often comes at a cost, but it lies in our hands to direct and secure AI technology to help and not destroy lives.

readmore

Archaeology Travel

I had the chance this year to meetup with my client Thomas Dowson from “Archaeology Travel Media” at the Travel Innovation Summit in Seville.

Over the past 2 years we have been revamping all the content from archaeology-travel.com and integrated a sophisticated travel itinerary builder system into the mix. We are almost feature complete and are currently fine-tuning the system. New explorers are welcome to signup and testdrive our set of unique features.

It was so nice to finally meet the whole team in person and celebrate what we have accomplished together so far.

What is Archaeology Travel all about?

Directly taken from the front-page :)

EXPLORE THE WORLD’S PASTS WITH ARCHAEOLOGY TRAVEL GUIDES, CRAFTED BY EXPERIENCED ARCHAEOLOGISTS & HISTORIANS

Whatever your preferred style of travel, budget or luxury, backpacker or hand luggage only, slow or adventure, if you are interested in archaeology, history and art this is an online travel guide just for you.

Here you will find ideas for where to go, what sites, monuments, museums and art galleries to see, as well as information and tips on how to get there and what tickets to buy.

Our destination and thematic guides are designed to assist you to find and/or create adventures in archaeology and history that suit you, be it a bucket list trip or visiting a hidden gem nearby.”

More Details
About
Mission & Vision
Code of Ethics

What is next?

We are constantly expanding our set of curated destinations, locations and POIs. Our plan is it, to make it even easier to find unique places for your next travel experience.

We are also working on partnerships to enhance travel options and offer a even broader variety of additional content.

Looking forward to all the things to come, as well as to the continued exceptional collaboration between all team members.

Lets continue to help bring the world’s past to the future :)

readmore

Chatbots

Typebot is a free and open-source platform that lets you create conversational apps and forms, which can be embedded anywhere on your website or mobile apps.

It provides various features like text, image, and video bubble messages, input fields for various data types, native integrations, conditional branching, URL redirections, beautiful animations, and customizable themes.

You can embed it as a container, popup, or chat bubble with the native JS library and get in-depth analytics of the results in real-time.

Typebot can be used in the Cloud (free tier available) or installed via Docker.

It drag & drop builder makes it amazingly fast to build your personal chatbot. It provides a ton of integrations (Google Analytics, Google Sheets, OpenAI, Zapier, Webhooks, Email, Chatwood) out of the box or you can connect other things via your own API Requests.

I have been a longtime user of RocketChat and build my own chatbots using Hubot or BotPress.

RocketChat
“Rocket.Chat is a free and open-source team chat platform that enables real-time communication between team members. It provides team messaging, audio and video conferencing, screen sharing, file sharing, and other collaboration features. It offers a familiar chat interface that makes it easy for teams to stay connected and collaborate effectively.

Rocket.Chat is highly customizable and extensible, with a large number of community-contributed plugins and integrations. It supports multiple platforms, including web, desktop, and mobile, making it accessible to teams regardless of their device or location. It also offers end-to-end encryption for secure communication, and compliance with various industry regulations such as HIPAA, GDPR, and others.”

Hubot
“Hubot is a popular open-source chatbot framework developed by GitHub. It allows developers to build and deploy custom chatbots that can automate tasks, integrate with various APIs, and enhance team communication and collaboration.

Hubot is built on Node.js and can be extended with various plugins and scripts, making it highly customizable and flexible. It comes with many built-in plugins, including integration with popular chat platforms like Slack, HipChat, and Campfire, as well as various APIs like Twitter, GitHub, and Google.

Hubot enables developers to build chatbots that can perform various tasks, such as scheduling meetings, deploying code, fetching data, and responding to user queries. It also provides a command-line interface for managing and interacting with the chatbot.”

Botpress
“Botpress is an open-source platform that allows you to build and deploy chatbots and conversational interfaces. It provides a user-friendly interface and a visual flow editor that allows you to create and manage your chatbot’s conversation flow. Botpress comes with many pre-built components and integrations, including natural language processing (NLP) engines, chat widgets, and webhooks, making it easy to build complex chatbots with advanced functionality.

Botpress offers features like machine learning-based NLP, conversational analytics, multi-language support, customizable themes, and a modular architecture that allows you to add custom functionalities with ease. It supports multiple channels, including web, Facebook Messenger, Slack, and WhatsApp, allowing you to reach your audience wherever they are.”

Typebot rocks

Typebot beats Botpress and could be the reason why I might be skipping RocketChat as well. Not sure I want to build an integration for RocketChat myself.

Setup

portalZINE NMN | Development meets Creativity | typebot

Base YAML Docker-Compose File to get you started. This will setup a Postgres instance, the typebot builder and typebot viewer.

Github
Docker Setup

readmore

THE QUESTION

A while back a potential customer asked me, if it is possible to restructure a WordPress Multisite setup and WPML with a more simplified and custom url structure?


THE ROUGH IDEA

1 . BASE.website web.site (with possibly different languages)

web.site/de/
web.site/en/

2. SUB.website web.site/nl-nl/

Languages would normally be added like this:

web.site/nl-nl/de/
web.site/nl-nl/en/

The customer wanted it to be restructured / simplified like this:

web.site/de-nl/
web.site/en-nl/

This basically mimics the structure of a single WPML website with custom languages, but with all the benefits of a multisite.

THE SOLUTION

This is nothing that WPML or WordPress Multisite provides out of the box.
I built a prototype setup to make it work.

Not something that I would propose for anyone, as it requires a lot of tweaks for anything that handles dynamic links (plugins, hooks, core systems, page.builder …)

Its doable :)

BASIC URL HANDLING

One thing that needs to be tweaked globally, is the mapping of the new url structure.

So web.site/nl-nl/en/ needs to become web.site/en-nl/

This needs to be handled on the server side, by proxying the original to the new structure.
This can be easily done using Apache or NGINX.

With that web.site/nl-nl/en/ will be proxied to web.site/en-nl/, but any core navigation will not work yet.

This is the fastest solution that I came up with, within the hour I gave myself ;)

There surely are other options, like the core rewrites / restructuring of the core shorturl handling. But these approaches might break things in far more areas.

Using the proxy approach, keeps the core as it is. The solution needs to be as simple as possible, allowing to maintain it in the future :)

HOOKS TO THE RESCUE

Just for the basic setup a couple of hooks are required to make this work, more might be needed depending on the plugins in use.

Here a couple of examples ….
WordPress site_url

WordPress Nav Links

WPML

Rankmath

This will not cover every angle, but will give you a starting point! I love my puzzles and there always is a viable solution :)

Need something similar … get in touch!

Happy coding …

readmore

Ich hatte dieses Jahr die Möglichkeit, über meinen Kunden TYPEMYKNIFE®, an der “Nacht der Sterne” in Stuttgart, im Mercedes-Benz Museum, teilzunehmen. Auf der Gala kommen mehr als 800 Gäste aus Gastronomie, Hotellerie, Politik, Kultur und Wirtschaft zusammen. 

NACHT DER STERNE / ahgz Sterne Award 2022

portalZINE NMN | Development meets Creativity | nacht der sterne 2022 typemyknife 1
Felicitas Then – Anmoderation
portalZINE NMN | Development meets Creativity | nacht der sterne 2022 typemyknife 3
Verleihung + Food

Es war ein klasse Abend, auf dem nicht nur die Spitzenköche aus Deutschland, der Schweiz, Südtirol und Österreich ausgezeichnet wurden, sondern diese auch Live zeigen konnten was sie so können.

Veranstaltet wird die Party von der Allgemeinen Hotel- und Gastronomie-Zeitung (Ahgz) und Burg Staufeneck / Rolf Straubinger.

Moderatorin des Abends waren die Journalistin und Fernsehköchin Felicitas Then und Rolf Westermann von der ahgz-Chefredaktion.

Informationen zum Award, den Methoden und Siegern findet man hier.

TYPEMYKNIFE®

portalZINE NMN | Development meets Creativity | nacht der sterne 2022 typemyknife 2
portalZINE NMN | Development meets Creativity | nacht der sterne 2022 typemyknife 4
TYPEMYKNIFE® – Im Gespräch

TYPEMYKNIFE® hat Vorort an einem Stand eine kleine Auswahl seiner Küchenmesser, die über den 3D Gravur Konfigurator vorbereitet und graviert wurden, präsentiert. Dadurch hatten Gäste die Möglichkeit, die gravierten Küchenmesser einmal persönlich zu entdecken und die Qualität zu bestaunen.

portalZINE NMN | Development meets Creativity | typemyknife portalzine
Glamour Shot / Kartentausch LOL

Die fast 1400 km Rundreise aus dem Norden hat sich gelohnt. Es ist immer schön Kunden mal nicht nur virtuell zu treffen, besonders wenn die Distanz so groß ist. Bei der Entfernung trifft man sich nicht immer mal kurz auf einen Kaffee oder Gin-Tonic :)

Gruß an TYPEMYKNIFE® / Schwäbisch Gmünd / Stuttgart

readmore

A grid mystery

Somehow the Grids Layout Builder for Gutenberg ist not pushing styles to the footer within a Timber theme. At least not for me.

Somehow the wp_footer action within the Grids plugin is not being executed and no grid styles are added to the footer. I am not getting any errors, but will have to investigate some more in the near future ;)

Update: It is related to the get_footer action within Grids. get_footer is used to load the appropriate footer template file, which I am not using in my Timber theme ;) So the workaround below is perfect.

Quick patch

Here a quick workaround for the theme functions.php, that does the trick for now.

Grids for WordPress, what is that?

“A layout builder is a tool that helps you creating visual structures in your page, from a simple layout made by adjacent columns, to more complex compositions.

Grids is entirely based on the WordPress block editor, which means that you’ll be able to use it together with the myriad of content blocks that developers and designers from all around the World are creating.

With Grids, we’re bringing a visual structure to the content written with the WordPress Block Editor.”
Grids: Layout builder for WordPress / Documentation / CSS-Tricks Article / A Complete Guide to Grid

Grids Example

This is using the Grids Plugin and only took a couple of minutes! I am so glad IE 11 is soon completely unsupported ;) and we can start using CSS Grid freely. Well we can, as nobody should be using IE 11 anymore!!!

On mobile its just stacked rows :)

CSS GRID: SOLVING PUZZLES

COLUMN 1
I love to solve puzzles and find epic solutions :)

CSS Grid Layout excels at dividing a page into major regions or defining the relationship in terms of size, position, and layer, between parts of a control built from HTML primitives.

Like tables, grid layout enables an author to align elements into columns and rows. However, many more layouts are either possible or easier with CSS grid than they were with tables.

LOVE IT!

COLUMN 2
We all love magazine-style layouts and we are getting closer to getting that :) The magic is in the detail.

For example, a grid container’s child elements could position themselves so they actually overlap another layer, similar to CSS positioned elements.

FLEX OR GRID
The basic difference between CSS Grid Layout and CSS Flexbox Layout is that flexbox was designed for layout in one dimension – either a row or a column.

Grid was designed for two-dimensional layout – rows, and columns at the same time.

COLUMN 3
I only styled the Desktop size, as a little preview. Added Custom CSS to switch areas for tablet sizes.

  • The grid properties are supported in all modern browsers.
  • An HTML element becomes a grid container when its display property is set to grid or inline-grid.
  • All direct children of the grid container automatically become grid items.
  • The spaces between each column/row are called gaps.

LET THE BATTLE BEGIN!

STAR TREK IPSUM

Unidentified vessel travelling at sub warp speed, bearing 235.7. Fluctuations in energy readings from it, Captain. All transporters off. A strange set-up, but I’d say the graviton generator is depolarized.

The dark colourings of the scrapes are the leavings of natural rubber, a type of non-conductive sole used by researchers experimenting with electricity. The molecules must have been partly de-phased by the anyon beam.

VS

STAR WARS IPSUM

Lucas ipsum dolor sit amet c-3po solo bothan qui-gon darth solo darth dantooine dagobah mustafar. Fett solo yoda r2-d2 kit obi-wan hutt amidala kenobi. Jade leia gonk lobot ahsoka darth jade skywalker organa. Utapau mara owen darth darth yavin.

Lando baba wedge darth solo skywalker ben fett. Fisto wookiee bothan antilles antilles luke kenobi. Yavin naboo kenobi jinn calamari antilles. Organa jabba skywalker gamorrean ackbar. Windu skywalker kit skywalker. Dantooine dantooine moff leia dantooine wicket amidala.

Things missing!

This is almost a perfect Gutenberg plugin and a reduced / simplified integration of CSS grid. I really think that this plugin makes Gutenberg 1000times better.

There is so much repetition of Gutenberg blocks out there, while this really addresses something new and nails it :)

There are still things missing, that are critical and would make this even more useful for certain websites:

  1. Allow to name grid areas. Would allow faster manual css modifications.
  2. Allow grid templates per breakpoint. This would make it perfectly usable in a responsive world. This is by far the most important issue!
  3. Some additional styling options would be nice, but nothing major. (Allow to adjust the grid gap …)
  4. Export the inline styles to files, that would leverage caching far better.
  5. Custom CSS integration, which allows to centralize Custom CSS for areas and sections.


Enjoy coding …

readmore

The issue

WACOM is known for stopping support for their pen tablets after some years, especially for the cheaper consumer pen tablets.

The WACOM Bamboo (in my case the CTH-461) is a pretty popular pen tablet and the latest driver wont install natively under Windows 10. Really frustrating for hardware that is still in perfect shape.

The solution

  1. Download latest available driver from Wacom
  2. Do not install the software directly yet.
  3. Right click on the downloaded file and choose “Properties”
  4. Click Compatibility on the popup
  5. Activate “Run this program in compatibility mode for”, choose Windows 8.
  6. Now you can install the driver & software.
  7. Enjoy

readmore