How many times have you gone through a check out to find that a barcode missing, mangled, or for some reason just refusing to scan?
Have you ever had to save your place in line while you run back to that aisle to find another product with a better barcode?
Even when barcodes do work, cashiers or customers have to spin their products around, looking for where they are located. In general, we spend way more time at the cash register than necessary.
Now, imagine a frictionless shopping experience.
These stores would be using a system that recognizes the product virtually instantly because it recognizes the entire product’s appearance. It’s free of the hassle of scanning barcodes and can be done in a couple seconds or less.
How could this efficiency be possible?
If companies implemented our algorithm into their systems used at checkout, we could make this frictionless checkout a reality.
Now we are going to address some concerns companies may have about adopting this algorithm.
It literally takes only a couple of minutes to create a database of 1,000 reference images. This approach is unlike other image searching technology that uses Machine Learning or SIFT methods. You will never have to train and re-train complex models like you would have to with our competition.
Object and product recognition through DejaVuAI is fast, reliable and incredibly simple to use. All you need is a handful of images taken at different angles for any product you want to find. Take a few reference images of each object with your phone or even with a cheap webcam and put images of each object in its own subfolder. No image processing is required. On average, 12-18 images per object is enough for reliable recognition.
Import them into the DejaVuAI database and begin recognizing your products.
There is no need to train an artificial neural network because there is none. If you need to recognize a new product, snap a few photos and add it to the database just as you did with the other images, which can be done in a matter of seconds.
Our technology is the most energy efficient solution on the market for image searching and product recognition because it can be run on low-power (and thus low energy consumption) devices.
Our solution is extremely cost effective.
We recognize products reliably and efficiently on a $1,000 PC. Because of this, there is no need to train a neural network or do the inference – we just search reference images (and we do not hide this fact). There is no expensive hardware needed, unlike other reverse image search technology which use inefficient and expensive ML or AI.
We have no problems scaling this system to support millions and even billions of products.
Our algorithm is extremely accurate, knocking out the competition when it comes to the results it brings to the table. Not only will it confidently identify clear images of products, but DejaVuAI will do something that no other technology can- it can find and match the reference images to images searched even when there are major distortions. There can be artifacts due to quality of the camera and lighting, the product could be turned or rotated, and there could even be other objects blocking part of the product you are trying to find, and DejaVuAI can still match it easily.
No matter whether you manufacture automated assembly lines, AR/MR/XR headsets, Point of Sale systems for frictionless shopping, or target recognition and guidance systems, you need our powerful, reliable and efficient DejaVuAI reverse image search and object recognition technology.
If you would like to see a demo of our DejaVuAI in action recognizing products, we invite you to watch the following video demonstration:
