#eyetracker

Eyetracking o Seguimiento Ocular

Los eye trackers funcionan captando el reflejo de luz infrarroja, invisible para los humanos, en el ojo, y calculando entonces el punto de fijación mediante complejos algoritmos matemáticos. Una vez determinado el punto de fijación, se puede visualizar o mostrar en la pantalla de un equipo. Existen 2 versiones, un eyetracking fijo que se pone en una pantalla para analizar normalmente campañas y…

El Eyetracking como centro del Diseño

A través de nuestros estudios hemos determinado el tiempo máximo que una persona permanece frente a una pieza de publicidad gráfica, son 3,2 segundos para uno físico y 1,8 segundos para digital. No es difícil para nosotros determinar cuánta información puede consumir una persona expuesta en ese tiempo, y esto se basa netamente en el funcionamiento del ojo humano, el ojo humano es como una cámara,…

Eye tracking is used to capture the gaze points as an observer performs a task. Eye tracking hardware produces an output video showing where a subject is looking, the subject’s point-of-regard (POR)/ Gaze Points, for each frame of a video of the scene. These gaze points have to be tagged with the corresponding image/object in all the eye tracking videos for all respondents involved in the market research. It’s only after the coding of these videos that we can analyze them for consumer engagement insights.

Manual Coding for Eye Tracking

To manually code gaze data, a person watches a composite video and codes, or scores, information about gaze behaviour. These gaze points have to be highlighted and marked for every frame of the video. Depending on the length of time during which gaze data (length of respondent video) is being collected and the sampling rate of the eye tracker, the amount of gaze data can be overwhelming. To analyze all the relevant gaze data points and to generate heat maps can take weeks and is an extremely tedious exercise. This process can be very time-consuming, cumbersome and subjective (the difference in perspective).

We must also account for track losses when analysing eye tracking data.

A track loss occurs when the eye tracking software fails to properly detect the pupil centres requiring the coder to find the start and end of every fixation (which requires the coder to navigate the frames one at a time around the vicinity of the desired frames).

Generating a gaze plot is the first step in analyzing the eye tracking videos and It can be complex for many first time/novice users. The platform/software offered by current companies can be disconcerting to many researchers. The agonizing process of gaze plot mapping can lead to inefficiency, affecting the final output.

SMART GAZE — AI Powered Coding Solution for Eye Tracking

Smart Gaze is an eye-tracking software which leverages its AI capabilities to simplify this coding process by automation. As a result, labour hours are considerably reduced, which allow professionals to work on other high priority jobs. AI-based coding has counted upon to have equal accuracy, if not better. The system uses historical data to determine gaze/fixations in the videos, which leaves no points of contention over points of relevance. There are no biases involved in the coding process and thus eliminating chances of human errors. Even with the high sampling rate, the accuracy of the output is not affected. It can code thousands of videos in a couple of days time, providing enhanced scalability.

Smart Gaze uses eye tracking coding which is compatible with almost all eye tracking hardware and it provides flexibility to the test subject during video recordings as even low quality (high noise) videos can be analyzed with equal accuracy. It allows for significantly lower turnaround time due to automation. Due to its automatic coding capabilities, it greatly reduces the repetitive and tedious aspects of manual coding. It is the most Cost effective solution in the market. Most manual coding software asks for subscription/license fee. However, Smart Gaze coding solution is based on a pay-as-you-use model which will save companies significant capital. Overall, Smart Gaze is a smarter solution to manual coding which gives you better insights into your market research.

“Object Detection” is a branch of Computer Vision that deals with finding specific objects (like humans, RedBull Cans, cartons of RedBull Cans etc.) from an image. With this blog, we will make a case about why Object Detection can be a game changer for market researchers.

Computer Vision is the field that deals with empowering computers the ability to ‘see’ things like humans. Object Detection is a basic visual perception task and one of the key areas of applications of Computer Vision. It essentially deals with finding and locating specific objects within an image.

For detecting generic objects (like a car, person, table, tree) there are open-source and pre-trained models like Yolo available. However, if you want an algorithm to detect very specific objects (like a ‘small raw tomato’ or a ‘large ripe tomato’), you will need to train an object detection algorithm of your own.

Take this video for example. In this experiment, our team was able to perform AI shelf analysis and identify all the red bull cans while a shopper strolls through a retail store. This is just one way in which market research can leverage the power of object detection for better consumer insights.

Another very important application can be in the field of retail audits.

Conducting retail audits is necessary to remain competitive in an industry where multiple products and brands fight for the attention of consumers but even today retail audit is a manual-intensive process. Hundreds of field agents roam around retail stores collecting shelf data and after compilation, they share it with the management. It sounds straightforward and simple, but it is anything but simple. The time-lag between a collection of data from retail stores until a final analysis is substantial which leads to companies losing time to take corrective action. By the time they realize the problem and formulate the solution, it is too late and the companies lose out on sales.

Data accuracy is another concern for management teams. They have to rely solely on the wisdom and honesty of the field agents to collect accurate data.

In today’s environment what companies need is a complete, end‐to‐end retail audit solution that would enable them to improve their shelf presence in real time. It is imperative for companies to fill the time gap between data collection and final analysis. This is where Object Detection comes to the rescue!

Using object detection for retail audits, companies will be able to improve the timeliness and accuracy of reports that will enable their managers to take in time corrective action.

https://www.youtube.com/watch?v=whM-vxbmUbg

Eye-Tracking is another very interesting field of market research that leverages the power of object detection.

In the video, you can see how visual input from eye tracking glasses has been coded to give insights about consumer behaviour inside a store. Eye tracking empowered with object detection can have a myriad of applications in market research such as Pack Testing, Advertisements, consumer behaviour analysis and many more.

In summary Object detection can help companies get important information about a number of key areas such as Actual requirement of a customer, Customer behaviour, his needs, habits, preferences, Main motivators and barriers at retail points, Optimize the point of sale (POS), Role of displays to the customers and many more.

We at Karna AI are leveraging the power of AI to bring object detection to practical use by integrating it with our market research solutions such as Smart Gaze for Eye-Tracking and Shelf Watch, our Shelf-Monitoring solution.