Take your A/B testing of your image or video advertisement to the next level. This will allow you to predict visual saliency of an image or video With computer vision and machine learning the engine lets you know what your customers will see, and what they will miss. and what part of ad will grab your consumers attention and optimize your next marketing campaign.
It shows the regions of the image/video that are most salient, using a heat map color coding from least salient regions (blue), to yellow, to orange and then red for most salient. The higher the saliency the more likely these items are going to be automatically noticed. Parameters such as contrast, brightness, density, angles, movement, colour composition, faces and more are determinants of this process
VISUAL COMPLEXITY –
This is an estimate of the amount of information that is in the image/frame. A single coloured display will have zero complexity; a highly complex white-noised image will have maximum complexity. In general, images with higher complexity can lead to easier distraction and thus increase the likelihood that people will miss out on crucial information (e.g. a brand, price or other communication)
SALIENCY COMPLEXITY –
This is an index of how many "hot spots" there are at any one time, or how "busy" the heat map is in an image/frame. This is indeed an index of how many things that are going to draw on people"s attention. If you have crucial information presented in an image/frame, then a higher saliency complexity is related to an increased likelihood that attention is distracted and communication efforts go down.
NEUTRAL – STATIC – LOW – HIGH –
A summary of how much of the heat map is neutral (has no saliency), static (blue scores on heat map), low saliency (yellow and orange) and high saliency (red). This indicates the distribution of saliency scores. Optimally, you would probably like to have that the high saliency has a low score, given that the hot spot is on the actual information you are conveying
COLOR LOAD AND BRIGHTNESS –
This summarises the degree of redness, greenness and blueness + brightness there is in the image. There are no recommended levels here, but the values are interesting parameters if you are expecting that an ad should have a very particular colour profile. It also gives you a good way of tinkering with different colour profiles in your visual materials.
VIDEO ANALYSIS –
This method allows you to see how visual saliency and the different complexity metrics vary over time. Visual complexity affects how well information is processed, and whether recipients are over flooded with information. Knowing when complexity is “just right” is essential for product and brand placement
Provides several benefits that outperform traditional methods:
• Fast : analysis results in hours, instead of weeks to months.
• Scalable : 1 image or video or a thousand images is all the same.
• Reliable : scientifically validated to predict over 85% of eye-tracking results.
• Usability : works by uploading an image to a dedicated folder, results provided in minutes to hours.
• Inexpensive : costs are a fraction of eye tracking and other traditional research methods.