ImageSleuth

Frequently Asked Questions

ImageSleuth is a web service that harvests demographic and other information from peoples’ faces.
ImageSleuth utilizes robust, high speed machine learning techniques to rapidly analyze images. From the user’s perspective it could not be easier: send us an image and we will return a JSON object with the results of the analysis.
ImageSleuth recovers details like age, race, and gender as well as hair and eye color, facial hair, eyewear, hats, and more. The service also recovers other information, like the pose of the face and whether the subject is looking at the camera.
ImageSleuth performance varies with the size of the images and the available bandwidth. We have been able to process tens of thousands of medium sized photos in an hour on a single multicore server. ImageSleuth is hosted on AWS, which means we have the ability to scale to meet any user’s capacity.
ImageSleuth derives its feature assignments by building composite faces from tens of thousands of individual subjects. We are thus able to harvest features without doing any form of face recognition, which preserves the anonymity of the subjects.
The results of the analysis are provided in JSON objects, which make it particularly easy to reuse in other applications.
Yes! You can use the API provided to integrate our service into your own web service in a matter of hours. We also provide custom integration services for customers who require them.
Yes! ImageSleuth not only provides results, but also provides a metric representing the degree of certainty in its outcomes, which will inform customers of the quality and certainty of the outcomes it produces.
Yes. Our modular design allows for the rapid addition of custom attributes, which can be up and running in as little as a few weeks.

You log into your account and are provided a secret key and secret token. You should be able to generate several keys and tokens pairs to distribute to employees.

First, you obtain an account and purchase a package. You will be provided with authentication information and a URL to access the service. Next, you download the ImageSleuth command line client. You then use the client to process directories of images. The client will produce JSON files for each of the images in the directory, each sharing the name of the photo from which it was derived.

Each ImageSleuth plan provides the customer with a specified number of image processing tokens. Each image processing token enables the user to process a single image. The larger the plan, the lower the cost per image token.
ImageSleuth provides an auto-renew functionality that automatically purchases additional tokens on the customer’s specified plan level. For example, if a customer purchased a 1000-image plan then the auto-renew would purchase another 1000-image plan once the number of remaining tokens fell below a percentage of the total number in the plan.

You pay in advance for the plan you want and the rate in which you use your plan is entirely up to you. We have an auto renew feature that insures that you never run out of images.