Getting Started Video Tutorials

Welcome to Feng-GUI dashboard, the most powerful predictive eye-tracking software in the market. Feng-GUI enables you to get instant and unbiased visual feedback at fraction of the cost and time of traditional eye-tracking. Thousands of professionals across all media channels are using Feng-GUI to test and optimize websites, advertisements, videos, print media, shelf layout, signage, and packaging design.

To analyze a new visual, the first thing to do is to Upload an image file of that visual stimuli. Let’s upload for example a snapshot image of webpage. Click Upload - image to select an image or a video file. You can upload up to 30 files at once. Wait until the file is uploaded and appear at the Media Library.

Next, we select the file at the Media Library, and we will set a few analysis parameters. To teach Feng-GUI about the content of this input visual, we set the View Type of the file as Website. To Set the viewer distance from this visual, we set the View Distance as “Desktop Screen” The remain options below enables you to adjust the format of the created reports.

Now, we add Areas of Interests. An Area of Interest is an edited area where the analysis metrics are calculated. such as visibility score, time to first fixation, fixations count etc.

Let’s click and drag a few boxes to create AOIs surrounding specific areas inside the visual stimuli. Tagging these areas allows Feng-GUI analysis engine to output AOI statistics for the Fixations & Visual Features Table which will appear here after we analyze the image.

Now we are ready to Analyze the image. We click the Analyze button and wait few seconds for the analysis to finish and the reports to load.

Welcome to Feng-GUI dashboard. In the previous tutorial, we uploaded a web page image, and defined few analysis options, and created some areas of interests, and clicked Analyze. And now when the analysis is over, the reports are loaded in the dashboard. Lets go over the reports.

The first report is the original image. The next one is the attention heatmap report.

The Attention Heatmap report, displays the most attractive elements of the image for viewers in the form of "hot" and "cold" spots creating so called heat map. The report is displayed as a color gradient overlay on the original image. The heatmap highlights areas likely to receive attention within the first few seconds. Use the Heatmap report to Visualize which elements attract more attention than others.

The Opacity map report, tones down information that is not attractive, and visually displays what your viewers may perceive during the first few seconds of visual inspection. Use the Opacity report to Emphasize which areas are being perceived and which are being ignored.

The Gaze Plot report, visualizes the scan paths and order between elements inside the image. It displays movement sequence, order and duration of gaze fixation. The report is made of a series of short stops (also called fixations) and fast movements of an eye (also called saccades). Use the Gaze Plot report to Visually demonstrate fixations order and gaze paths.

The Areas of Interest report, reveals a score of the predicted probability that a person will look at the area along with detailed diagnostics to show why this region is likely to get attention. You can get more details about this area at the table below. Here are the areas of interests that we've created, and their visibility score, and more information that also appears in traditional eye tracking, such as Time to first fixation, Fixation length and we've also added more details about the ingredients of the visibility score and what are the visual primitives that combines these scores. This are powerful insights because you can use these values to understand how to improve or how to reduce the attention towards this area. so for example, the tagline here has only 11 percent intensity, so if we will increase the intensity we will eventually improve the visibility score of this area. You can use the areas of interest report to as well compare variation of another design and compare areas from one design to another design variation

The Aesthetics report, enables you to examine the emotional aspects of aesthetics. Based on large scale studies with millions ratings of visual appeal collected from nearly 10 thousand participants, we developed computational models that accurately measure the perceived visual complexity and excitingness.

More information about the analysis reports, how to interpret and improve the results, may be found in the online help and documentation pages.

In Feng-GUI help page, input image section, you'll see a list of tips and practices. and in this session, we'll take a look at several examples. https://feng-gui.com/help

To begin with, input names, use alphanumeric characters. Otherwise, Feng-GUI will translate the file name into a unique alphanumeric characters file name. So to avoid that, just use English characters and numbers and dash and that's it. And this way, Feng-GUI will preserve your original file name.

Dimensions, input dimensions. So you don't want the input to be too small and you don't want the input to be too high. We have the recommended input size for images and for videos. If we take a look here at an example of an image which is too small, you may see that the image is almost the size of the legend overlay that we have here. And here is an example of an image with the proper dimensions. If we compare them side by side, you can see that the results are slightly different. Avoid using too small dimensions of an image or too large. Too large will cause no negative impact, only a little bit about performance.

Quality of the image and the video. Use only high quality and lossless compression. So don't compress your video or images too much. Otherwise you will get artifacts that are changing the image. As a result will also change the analysis report. Lossless compression, PNG and MP4 with 100% quality.

Product and package images. You should prefer landscape and square images. And add at least 10% of blank margin around the product. In this example the product is covering the entire image, which is not so good. You should leave at least 10% of space around the product. This is a good example. this is a good example, the product has some white space around it and the analysis is better.

A few tips about website input. We do have the upload web address but you should avoid it and just use your own snapshot that you create by yourself because it gives you more control of how the input image looks like. So some examples. So this is a good example of a web page image. You see that only the content of the web page is being analyzed and this is what you want to take a snapshot of, only the content of the web page. This is a bad example because in this example the image analyzed also contains the the toolbar and some artifacts of the browser and you don't want to analyze them. Only the content of the web page should be analyzed. This is another bad example. Obviously we don't see the entire web page at once so don't take a snapshot of the entire web page and analyze it. If you want to analyze all the web page from top to bottom either take different snapshots of the page or more easy and better approach is to take a video of navigating the web page and browsing through the web page and analyze this video in Feng-GUI. This will enable you more smooth transition and also enable you to record and analyze animation if there are or videos if they are in the web page.

Always prefer images which are landscape and horizontal because this is how we see the world and avoid analyzing pictures which are vertical and portrait because this is not how we see things. If this is what you have if that's the only image that you have so you can analyze it but if you are taking the picture outdoor or indoor so prefer taking landscape images obviously.

That concludes our session about file inputs and best practices for file formats, dimensions and quality. See you in the next session.

today’s session is about creating areas of interest, or AOIs. This topic is crucial for understanding how users interact with visual content, be it websites, advertisements, or packaging. AOIs help us focus our analysis on specific regions of interest within a visual stimulus, allowing us to gather more meaningful data about how users engage with those areas.

What Are AOIs? Let’s start with the basics. Areas of Interest, or AOIs, are specific regions on a screen or within a visual field that you define for the purpose of analyzing eye-tracking data. By setting AOIs, you can determine where users look, how long they spend looking at different areas, and how their gaze transitions between these areas.

Why Are AOIs Important? AOIs are crucial for several reasons: *Focused Analysis:* They allow you to isolate and analyze specific elements of a visual stimulus, such as a button, image, or text block. *Quantitative Metrics:* AOIs provide quantitative data on metrics like gaze duration, fixations, and saccades within defined areas. *Improving Design:* By understanding which areas attract attention and which do not, you can make data-driven decisions to enhance advertisement effectiveness and website design.

Setting Up AOIs** Now, let’s discuss how to set up AOIs effectively. There are several steps involved: *Define Your Objective:* *Purpose:* Determine what you want to learn from the analysis reports? Are you interested in user attention, interaction, or navigation patterns? *Content:* Identify the key elements on your screen or visual stimulus that you want to analyze.

*Add AOIs:* In Feng-GUI, use the AOIs "Add Area" menu, to create AOIs on your visual content. Carefully place AOIs around elements of interest. Ensure they accurately cover the areas you want to study.

Assign clear and descriptive labels to each AOI. This helps in organizing and interpreting your data. Use consistent naming conventions and sizing for AOIs across similar studies to maintain clarity and comparability.

Fine-tune the size and position of AOIs as needed to ensure they capture the relevant content without overlapping or missing important areas. Make sure AOIs are large enough to capture user interactions but not so large that they overlap with other AOIs. Validate your AOIs by running "Analyze" and or reviewing the reports to ensure they are capturing the intended areas effectively.

Automatic creation of AOIs using object detection algorithms works by identifying key objects or elements in an image, such as faces, products, or text, using machine learning models. These models scan the image, draw bounding boxes around detected objects, and automatically designate these areas as AOIs.

Analyzing AOI Data** Once your AOIs are set up, you’ll want to analyze the data collected. Look at metrics such as: "Time To First Fixation" The time it takes for a viewer's gaze to land on an AOI. "Fixation Duration" How long users spend looking at each AOI. "Fixation Count" The number of times users look at an AOI. "Gaze Patterns" How gaze transitions occur between AOIs. These metrics can provide insights into user engagement and the effectiveness of different visual elements.

Visual Features provide a comprehensive breakdown of the specific elements within your design that influence user attention. This feature analyzes key visual characteristics, such as contrast, color, edges, and shapes, which play a vital role in guiding the viewer’s gaze. By understanding how these features impact the overall Visibility Score of the AOI, you can make informed decisions about adjustments to improve visibility and focus on critical areas.

That brings us to the end of our lecture on creating AOIs for eye-tracking studies. I hope this has given you a solid foundation on the topic. Let’s open the floor for any questions you might have. Understanding how to effectively set up and analyze AOIs is essential for leveraging eye-tracking data to its fullest potential.

Welcome to Feng-GUI’s AOI's Compare Feature, a powerful tool for analyzing and comparing Areas of Interest across multiple images in your project.

With this feature, you can select specific images and instantly compare AOI values through a detailed table. For each image, the comparison table provides scores on key metrics: Clear – the clarity of the content, Complex – the cognitive demand of visual elements, Focus – the concentration of viewer attention, Exciting – the engagement potential, and Visibility Score for each defined AOI.

To ensure accurate comparison, it's essential that all selected images have the same AOI names defined. To facilitate this, you can easily copy AOIs from one image and paste them onto others. Once pasted, simply adjust each AOI’s position and size to ensure they align accurately with the intended elements in each design. This step is key for producing reliable and meaningful comparisons.

This comparison report is particularly valuable when evaluating different variants of packaging, web layouts, or ad visual designs. By comparing these metrics across versions, you gain insights into which design elements capture attention most effectively.

The report can be easily copied to the clipboard or downloaded as a CSV file, providing a convenient way to integrate insights into your workflow.

With Feng-GUI’s AOI's Compare, selecting the most effective design to enhance visual impact and viewer engagement has never been easier.

Thank you for your attention. See you in the next session.

Welcome to Feng-GUI. In this tutorial, you’ll discover how to optimize your website layouts using predictive attention analysis and our built-in AI Assistant.
Beyond uncovering how visitors view your Call-to-Action, the assistant provides data-backed recommendations and automatically generates a brand-new, optimized design mockup based on those insights.
To get started, let’s open an image we’ve already uploaded to our Media Library.

Next, let's automatically identify key visual components.
Go to the AOIs menu and click Create Auto AOIs.
Feng-GUI's object detection will instantly scan your layout and generate bounding boxes for major elements, including our target button, cta-3.

Before running the analysis, adjust your Analysis Options. Set the View Type to Online and choose your desired View Duration.
Now, click Analyze. Within seconds, your attention reports and AOI metrics table are generated.

Looking at cta3, we notice it has a low Visibility Score and viewers take too long to notice it.
Open the AI Assistant panel and ask:
'How can I improve the Visibility score and Time to first fixation of cta3?'
The AI assistant evaluates the visual features and provides actionable recommendations to boost its prominence.

Recommendations such as: Simplify and Shorten the Button Text. Increase Visual Intensity. and Incorporate Visual Cues.

Now, let's turn those recommendations into an actual visual mockup.
Ask the assistant:
'Use the above recommendations to create a new design image.'
The AI generates a new design variant implementing higher contrast, cleaner whitespace, and optimized hierarchy.

The generated image was added to the Media Library.
To ensure a consistent, accurate comparison, go back to the original image and select AOI's - Copy Areas.
Select the new design image and click AOI's Paste Areas.
Adjust the bounding boxes if needed, then click Analyze.
We can see that the Visbility Score if cta3 has improved.
Finally, select both the original and updated images in your Media Library and click Compare.
In the Compare Report, you can immediately see the improvement—higher visibility, faster fixation time, and a more engaging Call-to-Action.

With Feng-GUI, you can eliminate the guesswork and predict exactly where your audience will focus in the first critical seconds of visiting your page.
By combining predictive eye-tracking with the power of the AI Assistant, you don't just spot visual weaknesses—you instantly receive actionable recommendations and redesigned mockups to fix them.

Our side-by-side comparison report allow you to validate every single design decision before launching, ensuring your key calls-to-action get noticed faster and convert better.
Start optimizing your digital designs with confidence today at Feng-GUI.com, and turn visual attention into measurable results.

In this video, we’ll demonstrate how to analyze visual attention in a real-world design using an AI assistant and simulated eye-tracking reports from Feng-GUI

Let’s take this Pepsi Ad as our example. It’s professionally designed and visually striking — but what exactly draws the viewer’s attention first?

We’ll start by uploading the original Ad to the AI assistant. We ask the assistant to: "Analyze the design and provide pros, cons, and suggestions for improvement."

Even without additional data, the assistant can already offer an initial analysis. Let’s review some of the assistant’s feedback.

"The visual gag—Pepsi “wearing” a Coca-Cola costume for Halloween—creates an unexpected twist. Neuroscience research shows that surprises pique attention and boost retention, making the message more likely to stick in viewers’ minds." "The playful rivalry taps into humor, triggering positive emotional responses and dopamine release, which make people more likely to share or remember the ad."

"Potential Brand Confusion. The Coca-Cola branding is visually dominant due to the cape’s bright red and large logo. "

"Suggestions for Improvement. Consider pulling the “cape” back more to reveal more of the Pepsi can and logo. "

The AI Assistant highlights key strengths in layout, contrast, and visual hierarchy. It also identifies potential areas of distraction and how to improve them.

Next, we send the Feng-GUI heatmap report to the AI Assistant and the assistant to update its feedback based on the heatmap. "Attached is an analysis report of the design. Please update your earlier feedback using this report. Revise the pros, cons, and suggestions based on the visual attention data shown in the report."

Now, the assistant refines its insights using real attention data — highlighting hot zones and gaze patterns. "The heatmap reveals that viewers’ attention is overwhelmingly focused on the front of the can, with hot spots evenly distributed between the “Coca-Cola” cape and the central “Pepsi” logo."

The Assistant confirms that the central area in the ad receives most of the attention. saying that "The focal point draws immediate attention to the most conceptually important area—the can and the costume. This ensures viewers engage with the key narrative of the "brand rivalry" joke."

Honestly, I was amazed by how smart and helpful the AI assistant is — especially after analyzing the heatmap. It’s like having a second set of expert eyes on your design. What’s even cooler? This is AI-helping-AI helping you. That’s the kind of creative collaboration that helps you make better design decisions — faster and with more confidence. This kind of tool is a real game changer for anyone working on visual content.

With the power of AI and simulated eye-tracking, you can evaluate any visual — from ads to websites — and discover what truly captures attention. Use this workflow to guide your design process, test your assumptions, and build visuals that connect.

thank you for your attention, see you in the next session

Understanding Feng-GUI’s AI and Predictive Analytics Algorithms

Welcome to this lecture on Feng-GUI's cutting-edge AI algorithms, which have been at the forefront of predictive analytics since 2007. Let’s break down how Feng-GUI works and why it’s so effective.

Core Technology and Validation: Feng-GUI’s AI is rooted in Neural Network and Machine Learning and has been continuously refined and tuned for over two decades. Our algorithm boasts a remarkable 92% correlation with traditional eye-tracking studies, a standard verified through a benchmark by the MIT university. This accuracy is possible because the model is trained on a very large datasets and is regularly optimized to maintain its high performance. The datasets contains millions of data points across tens of thousands eye-tracking experiments.

Bottom-Up Algorithms: Feng-GUI incorporates and neurobiological principles to mimic how the human brain and eyes process visual information. An example of a Bottom-Up Algorithm is the Features Integration Theory, proposed by Anne Treisman, which explains how the brain combines visual features like color, shape, and movement. Another example for Bottom-Up Algorithms are saliency and computational attention algorithms, we use them to analyze how attention is drawn to different visual elements.

Top-Down Algorithms: Feng-GUI also integrates Top-Down algorithms. For example, we use Face Detection to highlight areas where human faces are present, as faces are naturally attention-grabbing. Further, our AI can interpret Facial Expressions, detecting eight different emotions from joy to anger. Similarly, Text Detection which works with any font and language. We also use algorithms to assess Aesthetic Factors such as balance complexity and focus in a design. All of the above are aggregated as Emotion Detection, classifying behaviors into Approach vs. Withdraw responses, which indicate positive or negative reactions to visual content.

Machine Learning and Optimization: Our AI is driven by Machine Learning, which enhances object detection and recognition capabilities. This allows for automatic creation of Areas of Interest (AOIs), speeding up the design analysis process.

OpenAI Integration: In recent advancements, we’ve integrated with OpenAI. This integration allows us to generating insights and actionable recommendations based on your reports. By leveraging OpenAI’s AI models, we provide more than just attention predictions; we offer strategic recommendations for improving your designs.

Conclusion: Feng-GUI’s AI combines bottom-up and top-down algorithms, machine learning, and predictive analytics to deliver highly accurate visual attention analysis. With a solid foundation of neuroscience and continuous learning, we provide actionable insights to enhance your designs, making the entire process faster, smarter, and more effective.

Thank you for joining this session!